HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dewin, D. R. Articles by Gompels, U. A. Search for Related Content PubMed PubMed Citation Articles by Dewin, D. R. Articles by Gompels, U. A.

Identification and Characterization of U83A Viral Chemokine, a Broad and Potent -Chemokine Agonist for Human CCRs with Unique Selectivity and Inhibition by Spliced Isoform¹

Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, University of London, London, United Kingdom

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Leukotropic human herpesvirus 6 (HHV-6) establishes a persistent infection associated with inflammatory diseases and encodes chemokines that could chemoattract leukocytes for infection or inflammation. HHV-6 variant A encodes a distant chemokine homolog, U83A, and a polymorphism promoting a secreted form was identified. U83A and three N-terminal modifications were expressed and purified, and activities were compared with a spliced truncated isoform, U83A-Npep. U83A efficiently and potently induced calcium mobilization in cells expressing single human CCR1, CCR4, CCR6, or CCR8, with EC₅₀ values <10 nM. U83A also induced chemotaxis of Th2-like leukemic cells expressing CCR4 and CCR8. High-affinity binding, 0.4 nM, was demonstrated to CCR1 and CCR5 on monocytic/macrophage cells, and pretreatment with U83A or modified forms could block responses for endogenous ligands. U83A-Npep acted only as antagonist, efficiently blocking binding of CCL3 to CCR1 or CCR5 on differentiated monocytic/macrophage leukemic cells. Furthermore, CCL3 induction of calcium signaling via CCR1 and CCL1 induced chemotaxis via CCR8 in primary human leukocytes was inhibited. Thus, this blocking by the early expressed U83A-Npep could mediate immune evasion before finishing the replicative cycle. However, late in infection, when full-length U83A is made, chemoattraction of CCR1-, CCR4-, CCR5-, CCR6-, and CCR8-bearing monocytic/macrophage, dendritic, and T lymphocyte cells can facilitate dissemination via lytic and latent infection of these cells. This has further implications for neuroinflammatory diseases such as multiple sclerosis, where both cells bearing CCR1/CCR5 plus their ligands, as well as HHV-6A, have been linked. Applications also discussed include novel vaccines/immunotherapeutics for cancer and HIV as well as anti-inflammatories.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Human herpesvirus 6 (HHV-6)³ primary infection causes widespread febrile illness in primarily infants, with a minority developing exanthema subitum, a mild skin rash. The virus establishes a latent infection that can reactivate in adults, mainly in immunosuppressed patients, to cause pathology, including transplantation diseases, bone marrow suppression, encephalitis, and links with other neuroinflammatory disease (1). The virus infects and persists in cellular mediators of immunity and, interestingly, encodes chemokine receptors and ligands that could mediate their recruitment as well as associated inflammatory disease (2). Thus, these genes may be essential for virus dissemination in vivo as well as virulence determinants.

HHV-6 exists in at least two strain groups, HHV-6 variant A (HHV-6A) and HHV-6 variant B (HHV-6B). They are closely related, with most conserved genes showing an average of only 5% sequence differences. Greater divergence is shown for selected loci at the ends of the genomes and a few specific sites between conserved gene blocks (2, 3, 4). These variants are related to a smaller genome of HHV-7 forming the roseoloviruses and, together with the more distant human cytomegalovirus, form the -herpesvirus subgroup of the herpesvirus family, maintaining a conserved gene order and similarity in sites of latency, including monocytic/macrophage cell types. From PCR-based sequencing studies, HHV-6A and HHV-6B have differing geographic prevalence, (5), with HHV-6B dominant in children from the United States, Europe, and Japan, whereas HHV-6A appears as only a minor variant, except in African countries, where it appears as equally prevalent to HHV-6B (1, 6). However, exhaustive surveys have not been conducted using serological-specific reagents, given the close relationship between these viruses.

There are hotspots for variation between representatives of these virus genomes, and these may contribute to some cellular tropism and pathological differences that have been anecdotally reported. For example, only HHV-6A has been detected in skin biopsies, and HHV-6A has been increasingly implicated in cases of multiple sclerosis (MS) where careful genotyping and identification of active infections have been conducted (1, 7, 8, 9, 10, 11). These studies implicate either immune abnormalities in clearance of the virus or possible complications of rare primary adult infection with this variant, because in countries where this has been studied, HHV-6B is the predominant variant identified. Both HHV-6A and HHV-6B have cellular tropisms for CD4⁺ T lymphocytes, and both are neurotropic, although there may be differences in the exact site of latency, given the more disperse detection of HHV-6A where studies have been undertaken. Recent studies further support an enhanced neurotropism of HHV-6A strains, suggesting sequence differences may affect biology and pathogenesis (12, 13, 14). Interestingly, the chemokine encoded by HHV-6 is one of the few hypervariable genes, with up to 15% sequence differences between these strain variant groups and thus would be a major candidate for determining pathogenic differences.

Chemokines are main mediators of an inflammatory response and can control chemotaxis of leukocyte populations to an infectious center (15). In human CMV (HCMV), for example, the UL146 chemokine is specific for -chemokine receptors and can control dissemination of the virus in specifically chemoattracted neutrophils (16). There is another locus encoding a potential HCMV chemokine, UL131-128, and this also appears to affect endothelial, leukocyte, and dendritic cell (DC) tropism, in that passage in fibroblasts results in deletion or alteration at this locus (17, 18, 19). At least one transcript here encodes similarities to -chemokines, which can chemoattract monocytes, and a similar function in murine CMV, vMCK, has been shown essential for virus dissemination (20, 21, 22).

In HHV-6, there is a single chemokine gene, which is deleted in HHV-7, encoding the highly variable U83 (2, 23). Initial studies on HHV-6B U83 had shown chemotactic activity for monocytic THP-1 cells, which express CCR2 and selected others and is consistent with the role of monocytic cells as sites for latent infection (24). Later studies demonstrated U83B as an efficient selective CCR2 agonist although with low potency (25). In the studies presented here, the functions of the HHV-6A U83 chemokine gene are examined. In this study, in contrast, U83A shows broad together with highly potent -chemokine activity for CCR1, CCR4, CCR5, CCR6, and CCR8. These receptors are represented on an interesting array of cell types, including monocytic, dendritic, and skin-homing Th2 cell types, which may contribute to the wider dissemination of this HHV-6 variant. Further discussed is a possible role for U83 in the detection of HHV-6A in a subset of patients with MS, given the implications of CCR1, CCR5, and CCR6 cells and their ligands in MS as well as in animal models for this disease. This broad specificity also may have useful applications in other inflammatory diseases, particularly asthma, as well as vaccine studies, such as cancer vaccines where targeting immune cells to improve antigenicity is crucial. Also shown are initial studies on the peptide encoded by the truncated U83A splice variant (23). Interestingly, this peptide lacks the agonist properties but can still interact with receptors as a competitive inhibitor. Given the different kinetics of expression with the spliced form early or latent and the full-length only late after DNA replication or lytic, such differential expression can have marked biological and pathological consequences.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Construction of GST-U83 expression plasmids

The U83 gene was amplified by PCR from viral DNA using three different primer pairs to give rise to three different constructs. All three forward primers added a 5' BamHI site to the U83 sequence, and the reverse primer (R83INT) added a 3' EcoRI site to the sequence for directional ligation of the U83 gene into the pGEX-2T plasmid and transformed into Escherichia coli strains DH5, and then the protease-deficient strain BL21 (Amersham Biosciences). Primers DD3 and R83INT were used to make pU83GST (containing thrombin recognition site). Primers DD11 and R83INT were used to make pU83GSTEK. An enterokinase (EK) recognition site was inserted 5' to the chemokine sequence and the thrombin recognition site provided by the plasmid. Primers FXa and R83INT were used to make pU83GSTXa. A factor Xa recognition site was inserted 5' to the chemokine sequence and to the thrombin recognition site provided by the plasmid. The U83 insert was sequenced from these plasmids using primers DD7 and DD12. The following primer sequences were used: DD3, 5'-TTGGATCCTTTATATGTAGTTCCCCCGATG-3'; DD11, 5'-TCGGGATCCCGTGATGATGATGACAAATTTATATGTAGTTCCCCCGAT-3'; FXa, 5'-TCGGGATCCCGTATCGAAGGTCGTTTTATATGTAGTTCCCCCGAT-3'; R83INT, 5'-CTTCGAATTCTTTCATGATTCTTTGTCT-3'; DD7, 5'-CCGGGAGCTGCATGTGTCAGAGG-3'; and DD12, 5'-AACGTATTGAAGCTATCCCAC-3'.

Expression and purification of recombinant U83, native and modified forms, was as follows. U83A and N-terminally modified forms were purified using the GST system as described (26)(Amersham Biosciences). Briefly, a single colony for each plasmid transformed BL21 E. coli (pGEX-2T parent plasmid and the recombinant U83-containing plasmids pU83GST, pU83GSTEK, and pU83GSTXa) was picked, used to inoculate 10 ml of Luria-Bertani medium (containing 100 µg/ml ampicillin), grown overnight in a 37°C shaking incubator, then 5-ml inoculated to 500 ml of Luria-Bertani-ampicillin medium and cultured to 0.5 OD₆₀₀. A total of 0.1 mM isopropyl -D-thiogalactoside was added to induce expression for 3.5 h, and then bacteria were pelleted and resuspended in 16 ml of STE buffer (10 mM Tris-HCl, 150 mM NaCl, and 1 mM EDTA (pH 8.0)) at 4°C, and then brought to 1.5% N-lauroylsarcosine, followed by cytolysis by sonication. The sonicate was centrifuged, and clarified supernatant was filtered through 0.45- and then 0.22-µm polyvinylidene difluoride filters (Fisher Scientific) and added to 1 ml of glutathione-Sepharose 4B bead slurry in STE buffer with 1.5% lauroylsarcosine (Amersham Biosciences). After mixing at 4°C for 1 h, the suspension was poured into a 10-ml polypropylene column for batch chromatography (Pierce) and washed with 20 ml of STE buffer with 1.5% lauroylsarcosine. Beads were washed with 8 ml of buffer A (50 mM Tris-HCl, 150 mM NaCl, and 5 mM ATP (pH 8.0)), followed by 8 ml of buffer A containing 0.15 mg/ml denatured BL21 proteins. The BL21 protein wash was required to remove a 70-kDa copurifying bacterial protein, presumed to be the bacterial chaperone dnaK, which recognizes foreign (non-E. coli), partially folded, or misfolded proteins in E. coli as described (27, 28) (GST system; Amersham Biosciences). Beads were then washed with 20 ml of STE buffer (containing 1.5% lauroylsarcosine), and bound protein was eluted with 8 ml of STE, 1.5% lauroylsarcosine plus 10 mM reduced glutathione (pH 9.0). The eluate was concentrated by centrifugation in a Vivaspin 20 (Vivascience), and then dialysed against PBS using SpectraPor2.1 membranes (Spectrum Laboratories), with a 15-kDa cutoff. Quantitation was done by Bradford assay (Bio-Rad), followed by cleavage using thrombin to give GS-83A, or protease site-modified N-terminal variants GSRDDDDK-83A or GSRIEGR-83A, EK, or factor Xa sites, respectively. To derive native U83A, EK or factor Xa was used on the pU83GSTEK- or pU83GSTXa-derived proteins, respectively. A total of 1.2 mg of fusion protein and 60 U of thrombin (60 µl in PBS; Amersham Biosciences) in 2 ml of PBS was incubated at 4°C overnight. For EK digestions, 1.2 mg of fusion protein was mixed with 6.4 U of EKMax (Invitrogen Life Technologies) in 50 mM potassium phosphate, 500 mM NaCl, and 50% glycerol (pH 8.0) diluted in 2 ml of PBS with EKMax reaction buffer (50 mM Tris-HCl, 1 mM CaCl₂, 0.1% Tween 20 (pH 8.0)) and incubated at 4°C overnight. For factor Xa cleavage, 1.2 mg of fusion protein and 6.4 U of factor Xa (6.4 µl; Amersham Biosciences) in 2 ml of PBS was incubated at 4°C overnight. Digestions were followed by analyses by SDS-PAGE and/or separated by reversed-phase HPLC (RP-HPLC).

SDS-PAGE and Western blotting

GST samples were analyzed using tricine SDS sample buffer (450 mM Tris-HCl, 12% glycerol, 4% SDS, 0.0075% Coomassie brilliant blue, 0.0025% phenol red (pH 8.45), tricine running buffer (100 mM Tris base, 100 mM tricine, 0.1% SDS (pH 8.3)) and precast over 10% tricine gels (Invitrogen Life Technologies). Gels were stained using Coomassie brilliant blue or SilverXpress kit (Invitrogen Life Technologies).

For Western blots, GST samples were blotted from electrophoresed gels using the XCell-II blot module with the Novex XCell Surelock mini cell, both from Invitrogen Life Technologies. Gels were blotted for 1 h at 25 V constant in Tris-glycine transfer buffer (12 mM Tris base, 96 mM glycine (pH8.3)) supplemented with 10% methanol, then blocked in 5% milk (Marvel) in PBS, followed by a wash in PBS-T (0.1% Tween 20). Blots were then probed with an anti-GST primary Ab (catalog no. 27-4577-01; raised in goat; Amersham Biosciences) diluted at 1/10,000, followed by the secondary Ab (catalog no. V8051; donkey-anti-goat-HRP; Promega) also diluted at 1/10,000. Alternatively, blots were probed with rabbit anti-U83 peptide Ab (kindly donated by Dr. G. Campadelli-Fume, University of Bologna, Bologna, Italy), diluted 1/10,000 in PBS-T followed by the secondary Ab (catalog no. W4011; goat-anti-rabbit-HRP; Promega) also diluted at 1/10,000, then washed in PBS-T, and incubated with ECL reagent (ECL Plus kit; Amersham Biosciences). The blot was exposed to Hyperfilm ECL (Amersham Biosciences).

RP-HPLC isolation and purification

Fusion protein and cleaved native or modified U83A were purified by RP-HPLC using a Resource RPC 3-ml column (15-µm polystyrene/divinylbenzene beads; Amersham Biosciences) and AKTA Explorer (Amersham Biosciences). Protein was acidified to pH 2.5 with trifluoroacetic acid to 0.1% (v/v), 2 ml loaded onto the column at 2 ml/min, and eluted using a buffer gradient with fusion protein eluting in 50.4% acetonitrile. Fractions were lyophilized and stored at –20°C.

Protease-cleaved native U83A chemokine and U83A N-terminal peptide were eluted in 45% acetonitrile, and fractions were lyophilized and stored at –20°C. The GST fusion proteins purified by RP-HPLC were resuspended in PBS (endotoxin certified) and quantitated by Bradford assay using Bio-Rad protein detection reagent and BSA standards (29). Samples of the native U83 chemokine (and N-terminal variants) were resuspended in water and quantitated by UV spectroscopy at 280 nm, with extinction coefficients calculated for each U83A variant as described (30). Resuspended and renatured chemokines in PBS were stored with 0.1% BSA (fraction V, endotoxin certified; Sigma-Aldrich) at –20°C. Endotoxin testing using Limulus amebocyte lysate assay showed levels <1 U/µg, equivalent or lower than commercially supplied human chemokines.

Cell lines and culture

THP-1 monocyte cell line (derived from monocytic leukemia) (31), K562 human lymphoblasts (from chronic myelogenous leukemia) (32), and HUT-78 T lymphoblasts of inducer/helper phenotype (from cutaneous T lymphoma, Sézary patient) (American Type Culture Collection) (33), and U937 monocytic cells were all cultured in RPMI 1640 without phenol red (Invitrogen Life Technologies) supplemented with 10% FCS (heat inactivated; Invitrogen Life Technologies) with 2 mM L-glutamine and for THP-1 cell with 0.1 mM 2-ME. Cells were used in signaling assays at early log phase between 0.2 and 0.6 million cells/ml. K562 cells stably transformed with pCDNA3 plasmid or pCDNA3 plasmid-expressing HHV-6 U51 gene were cultured as described (34) with 750 µg/ml geneticin G418 (Sigma-Aldrich). COS-7 cells were transfected with pCDNA3-expressing CCR1 or CCR5 (University of Missouri-Rolla cDNA Resource Center, Rolla, MO) using Lipofectamine 2000 (Invitrogen Life Technologies) following the manufacturer’s protocol. Chinese hamster ovary cells (CHO) were stably transformed with CCR1, CCR2, CCR4, CCR6, CCR8, CXCR1, and CXCR2. The gene expressing human G proteins, G16, was added to CCR1, CXCR1, and CXCR2 cell lines, while human Gqi5 was added to CCR4, CCR6, and CCR8 (GlaxoSmithKline). Primary human PBMCs were purified using Histopaque-1077 (Sigma-Aldrich), with monocytic/macrophage chemokine receptor expression (CCR1, CCR5, and CCR8) screened by flow cytometry. Every cell line used throughout this study was certified Mycoplasma-free at all times.

Flow cytometry

Receptor expression was confirmed by FACS and signaling assays using mouse mAbs specific for human chemokine receptors and isotype controls (R&D Systems) and read on a FACScan (BD Biosciences) with plotting using CellQuest Pro software (BD Biosciences). Nonadherent cells were dislodged, and adherent CHO cells were released using 20 mM EDTA (pH 8.0). Cells were centrifuged and resuspended in PBS with 2% FCS, and then 200,000 cells were preincubated with either BSA or human IgG (THP-1). Next, cells were centrifuged and washed in PBS with 2% FCS, incubated with chemokine receptor Ab or isotype control for 30 min at room temperature, and then centrifuged. Finally, cells were washed in PBS with 2% FCS and resuspended before reading on the FACScan.

Human chemokines

Human chemokine ligands were supplied lyophilized (R&D Systems), resuspended, and renatured as for U83A in PBS 0.1% BSA, with dilutions stored at –20°C. These included CCL1/I309, CCL2/MCP-1, CCL3/MIP1, CCL4/MIP1, CCL5/RANTES, CCL11/Eotaxin, CCL17/TARC, CCL19/MIP-3, CCL20/MIP-3, and CXCL8/IL-8.

Flexstation calcium mobilization assay

Intracellular calcium mobilization assays were performed using a Flexstation 96-well plate reader (Molecular Devices) with the FLIPR Calcium Plus Fluo 3-AM assay kit (Molecular Devices) as described by the manufacturer. Briefly, a total of 0.8 x 10⁵ cells/well, CHO-, or receptor-expressing cells were seeded into black-sided clear-bottomed 96-well plates and incubated overnight at 37°C, 5% CO₂. Medium was aspirated off the cells in the 96-well plate, and the cells were washed with a fluorometric imaging plate reader (FLIPR) buffer (145 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, and 10 mM glucose (pH 7.4)), made by the addition of 8.47 g of NaCl, 0.37 g of KCl, 0.29 g of CaCl₂, 0.20 g of MgCl₂, 2.38 g of HEPES, and 1.8 g of glucose to a final volume of 1 liter of water. Supernatant was aspirated and replaced with complete calcium plus reagent A,B FLIPR buffer containing Fluo-3 and probenecid, then incubated at 37°C and 5% CO₂ for 1 h. The plates were first read in an "end-point" assay to determine equivalent loading, with excitation at 485 nm, emission at 525 nm, with a 515-nm cutoff, and then read in a "flex" assay with ligand addition after 20 s of equilibration, with a read time of 120 s. Intracellular calcium release was plotted as sharp increases in fluorescence immediately after ligand addition, which gradually returned back to the basal level of fluorescence. The SoftMax Pro software (Molecular Devices) was used to calculate average values for each set of triplicate values, as well as maximum-minimum values for each average concentration. These were plotted using GraFit 32 software (Erithacus Software), a four-parameter fit was then applied to the values, and EC₅₀ values were calculated.

Transwell chemotaxis assay

Cells were fluorescently labeled with calcein using 2.5 µl of resuspended 5 mg/ml calcein-AM/DMSO in 600 µl in PBS, and 560 µl of calcein-AM was added to 7 ml of cells at 2 x 10⁶ cells/ml in complete medium (a total of 1.4 x 10⁷ cells/ml per 96-well plate), followed by incubation in the dark at 37°C for 30 min. Then, 31 µl of PBS/0.1% BSA (carrier) or chemokines diluted in carrier (100–0.001 nM) were added to the bottom well on the 96-well chemotaxis plate (NeuroProbe). The polycarbonate filter (5-µm pore) was replaced, and 50 µl of labeled cells was added to the top side wells and the plate was incubated at 37°C and 5% CO₂ for 2 h. Cells were aspirated off the top of the chemotaxis plate and replaced with 50 µl of 20 mM EDTA. The EDTA was aspirated off and the plate was read in a Wallac Victor (2) 1420 multilabel counter (PerkinElmer). Plates were read with an excitation wavelength of 485 nm and an emission wavelength of 535 nm, with 0.1-s interval between reads. The readout of the assay in relative fluorescence units (RFUs) represents the number of cells migrated as also checked by reconstructions with dilutions of counted cells. Each set was performed in triplicate, and SDs were calculated. Background random migration using only carrier gave the chemotactic index (CI) value of 1. CI values thus represent a fold increase of chemokine directed over the random migration. All values were plotted using GraFit 32 software.

Chemokine binding assay

Binding assays were performed as described previously (34). PMA- or IFN--differentiated U937 plus CCR1- or CCR5-transfected COS-7 cells were first tested for CCR1 or CCR5 receptor expression by flow cytometry, then washed in RPMI 1640 and resuspended at 2.5 x 10⁶ cells/ml in binding medium (RPMI 1640, 0.1% BSA, and 20 mM HEPES (pH 7.4)) on ice. Assays in triplicate contained 2.5 x 10⁶ cells, 166 pM radiolabeled chemokine (specific activity, 2000 Ci/mM, ¹²⁵I-labeled (¹²⁵I-)CCL3/MIP-1, or CXCL8/IL-8; Amersham Biosciences), and diluted concentrations unlabeled competitor chemokines (U83A or U83A-Npep as above, with human chemokines; R&D Systems). After 2-h incubations on ice, cells were separated from the unbound chemokine by microcentrifugation though a phthalate oil cushion (1.5 parts dibutylphthalate to 1 part bis(2-ethylhexyl)phthalate) with bound radioactivity counted with a gamma counter. Data analyses used Prism 0.1.53 software (GraphPad).

	Results

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Expression, purification, and identification of native HHV-6 U83A

Expression of U83A was examined in HHV-6A strain U1102-infected JJHAN cells (clone of CD4⁺ Jurkat T leukemic cell line). Both DNA and RNA from cDNA using RT-PCR were sequenced using designated primers from the flanking noncoding regions in the primary genomic sequence (2). The results showed that strain U1102 was polymorphic in this region, and an extended version of the gene was identified from a 2-nt insertion upstream of the start codon leading to the use of an upstream start codon (U83A polymorphism at position 123,510 bp in the published genomic sequence, a TT insertion, giving an U83A open reading frame of 123,482–123,823, from the original of 123,528–123,821; genome update submitted to GenBank) (2). Sequencing studies of clinical material also suggested that this may be a polymorphic site (data not shown). Similar genomic variability has been observed in a homologous region in HCMV encoding a gene product with similarity to -chemokines (17). The extended U83A now encoded a signal sequence comparable to that of U83B, with both start codons initiating at the same position, giving a similar predicted N terminus as shown by N-terminal sequencing (24). Studies of strain variation show two major groups, variant A and variant B strains, diverging by 13% with up to 3% variation within each strain group (23). Based on the polymorphism results, both the full-length U83A as well as the N-terminal spliced version (23) were produced without the signal sequence (24) to correspond to native proteins using primers described in Materials and Methods. The N-terminal spliced peptide corresponding to the N-terminal half of the product was produced by chemical synthesis, and the full-length product together with N-terminal variants by production in E. coli. Both were purified using RP-HPLC. The mature full-length U83A was amplified by PCR as detailed in Materials and Methods and expressed as a GST fusion protein in E. coli. The protein was purified by binding to a glutathionine column, followed by elution as described (26). The gene was mutated to include a N-terminal protease recognition sequence for factor Xa or EK, which could cleave at this site resulting in native, mature U83A released from the GST moiety of the fusion protein. Thus, three clones were produced: one in gGEX2T (with thrombin site), one modified with Xa site, and one modified with EK site. Three versions were produced by thrombin cleavage using the thrombin cleavage site encoded in the pGEX2T vector, giving GS-U83A, GSRIEGR-U83A, and GSRDDDDK-U83A. Two native U83A forms were produced using either Xa or EK. These had similar properties, but Xa was more reliable to use; thus, this construct was chosen to produce further quantities native U83A. After cleavage, the proteins were purified using RP-HPLC using a gradient of 40–60% acetonitrile (Fig. 1A). The chemokines separated distinctly from the cleaved and partial cleavage products of the GST fusion as shown in a silver-stained SDS-PAGE (Fig. 1B). The synthesized spliced product, U83A-Npep, was similarly purified and eluted in the same fractions as U83A (data not shown). The purified chemokines were identified by Western blotting using a U83A peptide-specific polyclonal sera (Fig. 2, A and B) but were not recognized by GST-specific sera (C), although this was specific for the full-length products only because it was derived from a peptide covering the splice donor site (B). The N-terminal sequence was confirmed by Edman degradation, and sizes were similar to those produced by in vitro transcription translation of both the spliced and full-length versions of U83A (23); these also appeared as a doublet, probably from posttranslational modifications such as phosphorylation, because there are numerous sites for this modification in the coding sequence. The purified chemokines were dialysed against PBS, and then freeze-dried for storage. Chemokines were solubilized and refolded in PBS and stored at –20°C for <1 mo before assay.

View larger version (51K): [in this window] [in a new window]   FIGURE 1. Purification of U83A. A, Shown are traces (A280) from three separate purifications using RP-HPLC. B, Silver-stained SDS-PAGE of peak fractions shows the separation of native U83A in fractions 49 and 50 from uncleaved U83GST fusion protein and cleaved GST in fractions 56–60.

View larger version (33K): [in this window] [in a new window]   FIGURE 2. Identification of purified U83A and variants by Western blot. A, Silver-stained SDS-PAGE showing HPLC-purified U83A-Npep, U83A purified after EK cleavage, U83A(EK) (EK-U83-GST construct), or after Xa cleavage U83A(Xa) (Xa-U83A-GST construct), or N-terminal variants of U83A as indicated (GS-U83A, GSRIEGR-U83A, and GSRDDDDK-U83A) purified after thrombin digestion of constructs U83A-GST, Xa-U83-GST (factor Xa site N-terminal modified U83A), and EK-U83A-GST (EK site N-terminal modified U83A). MW are m.w. markers as indicated, and GST is the HPLC-purified GST moiety. B, Western blot results using anti-U83A intron peptide sera identified native full-length U83A and N-terminal variants. C, Control Western blot results using anti-GST sera confirm specificity and purity, identifying only the GST.

Both full-length U83A and U83A-Npep were tested in calcium mobilization assays for functional chemokine activity. Surprisingly, given the positive results with 100 nM U83B-Fc on THP-1 cells and CCR2-transfected L1.2 cells (24, 25), initial results with 100 nM U83A using THP-1 premonocytic cell lines with quantitation using individual cell image analyses only gave isolated responses from single cells in the population, whereas the CCR2 ligand CCL2 induced >95% of the cell population (data not shown). These were cells that were polarized to express CCR2 over CCR1 and CCR5 as shown by flow cytometry as well as assays using the CCR2 ligand CCL2 in calcium mobilization and chemotaxis assays. In contrast, the K562 pre-erythroid cells that transiently express CCR8 showed potent calcium mobilization of >95% of the population. To follow this further in a defined setting, we used CHO cells expressing individual chemokine receptors CCR1, CCR2, CCR4, CCR6, CCR8, CXCR1, and CXCR2. CCR3 also was assayed in CCR3-K562 cell lines using a scintillation proximity assay where no inhibition by U83A (0.1–50 nM) was found of ¹²⁵I-CCL11/eotaxin binding to CCR3 (data not shown). The results showed that full-length U83A could potently and efficiently induce calcium mobilization via CCR1, CCR4, CCR6, and CCR8, whereas the spliced variant, U83A-Npep, had no effect (Fig. 3). The EC₅₀ values for CCR1 and CCR8 were similar to those of the endogenous ligands, with U83A at 9.17 ± 0.2 and 6.42 ± 0.5 nM, respectively, and with CCL3 and CCL1 at 2.75 ± 0.9 and 3.87 ± 0.8 nM, respectively. In contrast, the EC₅₀ values to CCR4 and CCR6 at 8.96 ± 0.27 and 10.3 ± 0.8 nM were at least 10-fold less sensitive than those of the endogenous ligands at 0.24 ± 0.05 nM (CCL17) and 1.27 ± 0.03 nM (CCL20), respectively. Comparisons of activities to all the cell lines and the parental cell lines showed significant responses only to CCR1, CCR4, CCR6, and CCR8 (Fig. 4). There were marginal responses to CXCR1, but these were only at relatively high concentrations and diluted out at 10 nM to background levels. Although CCR5-CHO cells also were tested, receptor expression was too low to give consistent results. Responses to CCR2 and CXCR2 were indistinguishable from background levels on the CHO parental cells. A low activity to a hamster homolog of CCR8 was detected in responses to high concentrations of CCL1 (>100 nM), which was equivalent to these background levels. At 100 nM, U83A showed >100% activity, compared with endogenous ligand CCL3, whereas responses via CCR8, CCR4, and CCR6 were 85, 84, and 80%, respectively. At 10 nM, these responses dropped to 60, 65, 50, and 35%, respectively (Fig. 5). Similar results were found with the N-terminal variants, although the spliced truncated form, U83A-Npep, showed no activity at all concentrations tested (to 100 nM).

View larger version (37K): [in this window] [in a new window]   FIGURE 3. Induction of intracellular calcium mobilization using U83A and human endogenous chemokine ligands. CHO cell lines stably transfected with CCR1, CCR4, CCR6, and CCR8 and labeled with Fluo-3 are treated after 20 s with 50 nM U83A, U83A-Npep, or endogenous human chemokine ligands CCL3, CCL20, CCL17, and CCL1, respectively. RFU indicates relative levels of induction of calcium mobilization. EC50 values were calculated as indicated from measuring peak induction levels after stimulation with 100, 50, 10, 7.5, 2.5, and 1 nM concentrations of indicated U83A or human chemokines. Negative controls were the vehicle buffer without added chemokine.

View larger version (21K): [in this window] [in a new window]   FIGURE 4. Dose response of U83A induction of calcium mobilization in CCR1, CCR4, CCR6, and CCR8-CHO cell lines. RFUs and the results performed in triplicates with SD are shown. Activities are compared with background on CHO parent cells and low or negative activities on CXCR1, CXCR2, and CCR2-CHO cell lines. Activities at 1 or 100 nM U83A-Npep are equal to baseline levels with negative control vehicle buffer only.

View larger version (18K): [in this window] [in a new window]   FIGURE 5. U83A induction of calcium mobilization in CCR1, CCR4, CCR6, and CCR8-CHO cell lines, compared with reactivity to endogenous ligands. Expressed as percentage of maximal response of chemokine-induced calcium mobilization in CHO cells expressing cellular chemokine receptors indicated with CCL3/MIP-1 to CCR1, CCL22/MDC to CCR4, CCL20/MIP-3a to CCR6, CCL1/I309 to CCR8, CXCL8/IL-8 to CXCR1, CCL2/MCP1 to CCR2, and CXCL8/IL-8 to CXCR2. CCR2 and CXCR2 at U83A 100 nM are not plotted as though they have values significantly above the vehicle-only background they are indistinguishable from background responses on the CHO parent cell line (see Fig. 2). Baseline with vehicle-only values subtracted.

To investigate further functional responses to U83A, chemotaxis assays were performed with monocytic and T lymphocytic cell lines with defined chemotactic properties and chemokine receptor expression. Chemotaxis assays were first performed using the monocytic THP-1 cells given reactivity demonstrated with U83B-Fc fusion proteins (24). Later results of chemotaxis with CCR2-transfected L1.2 cells using native U83B synthesized form also were consistent because THP-1 cells can express CCR2 (25). In contrast to results with U83B, U83A (0.1–100 nM) induced no chemotaxis with THP-1 cells demonstrated to be polarized for expression of CCR2 by flow cytometry and calcium mobilization to CCR2 ligands, as well as CXCR4 (data not shown). This was consistent with the calcium mobilization data with U83A using CCR2-CHO cells or THP-1 cells, which was indistinguishable from background readings.

Given the reactivity to CCR4 and CCR8 CHO cell lines in calcium mobilization experiments, chemotaxis assays were performed on a clone of the HUT-78 T leukemic cell line, which had a Th2 cell phenotype expressing CCR4 and CCR8 (as demonstrated by FACS). Reactions with ligands CCL17, CCL22 (CCR4), and CCL1 (CCR8), respectively, resulted in efficient and potent chemotaxis of this Th2 cell line with maximal efficacy at 1, 10, and 1 nM, respectively (Fig. 6). The results showed that U83A also efficiently and potently chemoattracted this cell type. Ranges tested were from 0.0001 to 100 nM, with minimal efficacy at 0.1 nM. At 100 nM, the U83A response was 70% of the response of the endogenous ligands CCL17 (for CCR4) and CCL1 (for CCR8), with 45% of CCL22 (also for CCR4). Chemotaxis normally has a bell-shaped curve as shown for the endogenous ligands (Fig. 6), whereas for U83A, the curve was still rising at 100 nM; thus it is still possible there were more potent reactions at higher concentrations. Interestingly, using N-terminal variants of U83A with N-terminal extensions (encoding the protease recognition sites), similar results were demonstrated (Fig. 6), although most potent responses at 10 nM were shown by the unmodified U83A.

View larger version (28K): [in this window] [in a new window]   FIGURE 6. Chemotaxis of HUT-78 T lymphocyte cells showing responses to U83A and variants, compared with endogenous human chemokine ligands, the CCR4-specific CCL22 and CCL17, and the CCR8-specific CCL1. Also shown are the negative control of replicates of cells treated only with ligand-free buffer (vehicle). A dose response is indicated showing reactions with 10-fold serial dilutions of concentrations from 0.0001 to 100 nM. Error bars are shown for results repeated in triplicate.

To test the specificity of the U83A-induced chemotaxis of the HUT-78 cells, tests for antagonism were conducted by pretreatment with U83A before assay with the endogenous ligands using a chemotaxis assay. The results showed that pretreatment with 100 nM U83A or the N-terminal variants partially blocked chemotaxis using the maximal effective doses for CCL22 (10 nM) for CCR4 and CCL1 (1 nM) for CCR8. For antagonism of CCR4 responses, a trend for greater antagonism with larger N-terminal extensions was shown, with 25% blocking using pretreatment with GSRIEGR-U83A (Fig. 7A). For antagonism of the CCR8-mediated response, native U83A blocked 45% of CCL1-induced chemotaxis, whereas N-terminal extensions were not effective. Even the smallest modified form, GS-U83A, blocked only 10% of the endogenous ligand activity (Fig. 7B). Similar results were found in antagonism of calcium mobilization using a FLIPR-based assay with CHO-CCR4 and CHO-CCR8 cells (data not shown). Activity against CCL22-induced chemotaxis via CCR4 was observed only at the highest concentration assayed, 100 nM, whereas a dose-response effect could be shown for blocking only by native U83A against CCL1-induced chemotaxis via CCR8 (Fig. 8). In this study, the concentrations were tested from 0.0001 nM, and the minimal effective dose for antagonism was demonstrated from 1 nM.

View larger version (14K): [in this window] [in a new window]   FIGURE 7. Competition by U83A and variants of chemotaxis induced by human endogenous ligands CCL22 on CCR4 or CCL1 on CCR8. HUT-78 cells were preincubated for 30 min with vehicle only or 100 nM U83A, GS-U83A, GSRDDDDK-U83A, or GSRIEGR-U83A, then treated with maximally responsive endogenous chemokine concentrations of 10 nM CCL22 (for CCR4) shown in A, or 1 nM CCL1 (for CCR8) shown in B, as indicated. Endog, Endogenous. The results are presented as percent inhibition of cell migration as measured by RFUs.

View larger version (27K): [in this window] [in a new window]   FIGURE 8. Dose response of competition effect by U83A on chemotaxis induced by endogenous ligand (CCL1) for CCR8. Competition was shown from 1 nM. The N-terminal variants only showed competition at the highest concentration tested, 100 nM, against CCL17-induced chemotactic responses of CCR4 CHO cells (see Fig. 7) with a marginal effect from 100 nM GS-U83A on CCL1-induced CCR8 chemotaxis. Relative CI is indicated as shown (Fig. 6) against a dose response showing reactions with 10-fold serial dilutions of concentrations to 100 nM. Error bars are from results repeated in triplicate.

To investigate the mechanism of both agonist and competitor activities of U83A, binding assays were performed. First, COS-7 cells expressing CCR1 were used in binding displacement assays using ¹²⁵I-CCL3. The results showed efficient binding with an affinity similar to or higher than that of endogenous ligands at 0.4 nM (Fig. 9A). Because the CHO-CCR5 cells expressed levels of CCR5 that were too low or variable to perform functional assays, interactions were further investigated in this study using COS-7 expressing CCR5 cells in binding assays. The results showed very high-affinity binding at 0.06 nM, displaying a preference for CCR5 over CCR1 (Fig. 9A). Next, binding was investigated on a human monocytic/macrophage cell line, U937 cells induced by PMA, expressing CCR1 and predominantly CCR5. Results here showed similar high-affinity binding at 0.03 nM (Fig. 9A). Interestingly, in displacement assays using the U83A-Npep, there also was specific albeit lower binding with an affinity of 54 nM (Fig. 9B). Using COS-7-, CCR1-, or CCR5-expressing cells, 40 nM U83A-Npep also efficiently displaced 50–60% of ¹²⁵I-CCL3 binding, respectively, showing that either receptor on the monocytic/macrophage cells can be inhibited by the U83A-Npep pretreatment; whereas U83A-Npep could not displace binding of ¹²⁵I-CXCL/IL-8 on CXCR2-expressing THP-1 cells. Conversely, CXCL/IL-8 could not displace binding of CCL3 to its receptor CCR1 or CCR5 expressed on either COS-7 or the monocytic/macrophage cells, thus demonstrating the specificity of both assays.

View larger version (18K): [in this window] [in a new window]   FIGURE 9. U83A and U83A-Npep affinity for human chemokine receptors. A, COS-7 cells transfected with plasmids encoding human CCR1 () or CCR5 () were used to assess the affinity of U83A. A total of 2.5 x 107 cells/well were incubated with 125I-CCL3 (166 pM) for 2 h in the presence of increasing amounts of U83A. EC50 values are 3.9 ± SD 1.3 x 10–10 and 5.6 ± SD 1.8 x 10–11 M, respectively. Binding specificity was checked by control displacements induced by 40 nM CXCL8 on CCR1 () and CCR5 () cells. The total binding of radiolabeled ligand in the absence of any competing ligand was taken as 100%. Each point represents the mean ± SD for determinations performed in triplicate. B, Competition binding assay of U83A and U83A-Npep on U937 PMA-induced monocytic/macrophage cells. A total of 2.5 x 107 cells/well were incubated with 125I-CCL3 (166 pM) for 2 h in the presence of increasing amounts of U83A () or U83A-Npep (). EC50 values are 3.1 ± SD 1.2 x 10–11 and 5.4 ± SD 1.6 x 10–8 M, respectively. The total binding of radiolabeled ligand in the absence of any competing ligand was taken as 100%. Each point represent the mean ± SD for determinations performed in triplicate from at least two different experiments.

View larger version (19K): [in this window] [in a new window]   FIGURE 10. Antagonism by U83A and U83A-Npep of CCR1-mediated induction of calcium mobilization by U83A (GSRIEGR-U83) or endogenous ligand CCL3. CCR1-CHO cells were pretreated with 100 nM endogenous chemokine ligand CCL3, U83A, or U83A-Npep. Negative controls were the nonbinding CCL1 or the vehicle buffer without any chemokine. Then the pretreated cells were stimulated with the endogenous ligand CCL3 or the viral chemokine U83A as indicated. The maximum peak of calcium flux induced as measured in RFU was expressed as a percentage of inhibition of the response from pretreatment only with buffer. The results show CCL3 antagonizes induction of calcium mobilization by U83A or is competed out with pretreatment also with CCL3; U83A antagonizes induction of calcium mobilization by both U83A or CCL3; U83A-Npep antagonizes induction of calcium mobilization induced by U83A or CCL3. U83A-Npep blocks CCL3-induced responses as efficiently as pretreatment with the endogenous ligand. Negative control CCL1 does not antagonize calcium mobilization induced by CCL3 or significantly by U83A. All experiments were performed in triplicate, and error bars are indicated.

View larger version (12K): [in this window] [in a new window]   FIGURE 11. Inhibition of CCL1-induced chemotaxis by U83A-Npep via CCR8 on primary human PBMCs. PBMCs (105 cells/well) were added after pretreatment with either U83A-Npep (2.5 mM) or DMSO and left to migrate for 2 h at 37°C. Statistically significant differences (Mann-Whitney U test) in CIs, between CCL1 after pretreatment with DMSO alone or with U83A-Npep, are indicated as follows: 10 nM, p < 0.05; 1 nM, p < 0.01. The chemotactic responses, expressed as the mean CI ± SD, are derived from two independent experiments.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

These results show that, in functional assays using both calcium mobilization and chemotaxis, HHV-6 U83A can function as a potent, selective chemokine agonist with broad -chemokine reactivities to CCR1, CCR4, CCR6, and CCR8. Binding studies also show high-affinity binding for CCR1 and CCR5 at subnanomolar levels. The most potent responses were to CCR1, CCR5, and CCR8, which were similar to those found for the endogenous ligands. In comparisons using calcium mobilization assays, the sensitivity relative to the endogenous ligand interactions was CCR1 > CCR8 > CCR4 > CCR6. The relative potency of the responses was demonstrated by the minimal effective doses shown for CCR8 at 2.5 nM and at 7.5 nM for CCR1, CCR4, CCR6, and CCR8 in calcium mobilization assays and 0.1 nM in chemotaxis assays. Specificity was shown in blocking activities of endogenous ligands in calcium mobilization and chemotaxis assays. Responses were efficient showing 80–100%, compared with endogenous ligands. U83A showed very high-affinity binding to CCR5 at 0.06 nM, with high affinity to CCR1 at 0.4 nM, whereas U83A-Npep showed moderate affinity at 54 nM, indicating ligand binding in the N-terminal domain of the molecule with C-terminal regions enhancing binding.

We have shown previously that expression of U83 expressed early in infection is modulated by novel cellular splicing that results in the introduction of a stop codon, causing truncation of the chemokine gene after the first set of encoded conserved cysteines (23). The full-length version is only produced late in infection after DNA synthesis. Interestingly, the spliced version, U83A-Npep, showed no agonist activity, only antagonist activity as demonstrated for CCR1, CCR5, and CCR8 in binding, calcium mobilization, or chemotaxis assays; thus, it retains binding, albeit at a lower affinity, but does not signal, suggesting the C-terminal half of the molecule is required for intracellular signaling. This would be consistent with antagonist activity early in infection, which could act in protecting the infected cell from immune cell surveillance, as demonstrated here by chemotaxis inhibition, via pretreatment with the spliced variant, of primary human mononuclear cells expressing CCR8 found on monocyte/macrophages and T lymphocytes. Although late in infection after DNA replication and virion production, the agonist properties of the full-length chemokine can chemoattract similar cellular populations now for virus dissemination or latent infection.

Only minor differences were found using the N-terminal variants, GSR-, GSRDDDDK-, and GSRIEGR-U83A, compared with native U83A in assays of chemotaxis or calcium mobilization, whereas there were more marked differences in the antagonist activity. In this study, only native U83A was effective in blocking the CCL1-induced chemotaxis via CCR8, whereas only the larger N-terminal extensions showed some partial activity in blocking the CCL22-induced chemotaxis via CCR4. This suggests that the sites for interaction of viral and endogenous chemokines, particularly on CCR4, are overlapping rather than identical.

Although we have shown here that U83A appears to have a broad -chemokine specificity with high potency, recent studies on U83B encoded by HHV-6B (25) show a selective CCR2 activity with low potency. This could highlight some of the subtle cellular tropism differences that have been characterized between the strains using leukemic cell lines for cultivation (1). The results suggest that U83A could chemoattract a wider range of cell types for further dissemination. Of particular interest are the skin-homing T cell properties of CCR4 and CCR8 cells, and in vivo analyses of sites of persistent infection by PCR analyses of biopsy material showed frequent detection of HHV-6A strains at sites in the skin (35) and could explain the wider distribution of HHV-6A strains identified to date, for example, in lung and neuronal tissue (12, 13, 36). The CCR4/CCR8 phenotype of Th2 cells also could contribute to the well-defined cellular tropism of HHV-6 for mature CD4⁺ T lymphocytes (37) in chemoattracting this cellular population for lytic infection, whereas all of the reactive receptors are present on T lymphocytes. Monocytic/macrophage cell types have been described for both lytic and latent infection (38, 39). CCR2 is present on human monocytic cells, whereas CCR1, CCR5, and CCR8 on differentiated monocytic/macrophage cell types (40, 41). Reactivities with primary differentiated human monocytes were demonstrated by chemotaxis antagonism with the U83A spliced variant. As HHV-6B U83B can chemoattract CCR2-bearing monocytes, while HHV-6A U83A chemoattracts CCR1, CCR5, or CCR8-bearing differentiated monocytic/macrophages, this distinction could contribute to the differences in latency and cell tropism.

Comparisons to other human chemokines

HHV-6A U83A is novel as the only broad, potent -chemokine agonist. Currently, CCR4, CCR6, and CCR8 are restricted in human chemokine ligands, with only two defined for CCR4, CCL17 (TARC) and CCL22 (MDC); one for CCR6, CCL20 (MIP-3); and one for CCR8, CCL1 (I309). TARC is primarily for CCR4, to a lesser extent to CCR8 (42). CCR1 interacts with a wider range of ligands CCL3, CCL3L1, CCL5, CCL7, CCL14, CCL15, CCL16, CCL9/10, and CCL23. Although most of these are restricted to CCR1, the exceptions are CCL3 (CCR5), CCL5 (CCR3 and CCR5), CCL7 (CCR2 and CCR3), CCL14 (CCR5), and CCL16 (CCR2). CCR5 ligands also include CCL3, CCL4(MIP1B), CCL5(RANTES), and CCL8(MCP2). Thus, U83A has a unique combination in functional interactions among CCR1, CCR4, CCR5, CCR6, and CCR8. The cell types that these receptors are present on represent a distinct array of monocytic/macrophage (CCR1, CCR5, and CCR8), T lymphocytes (all), Th2 lymphocytes (CCR4 and CCR8), and immature DCs (CCR1 and CCR6), as well as reports on NK, eosinophil, and endothelial cells (CCR8), which all can be targets for lytic or latent infection by HHV-6. Thus, full-length U83A, made after DNA replication, could chemoattract these cell types, leading to dissemination of the virus, whereas U83A-Npep made before replication could mediate their blocking, thus avoiding innate and adaptive immunity directed by these cell types.

Comparisons with other viral chemokines

In the studies shown here, a minority of cDNAs from early passage virus encoded the complete U83A gene with an extended N-terminal region encoding a signal sequence competent for secretion comparable to that in U83B. Thus, in passaged virus, a small deletion interrupts U83A to give an internal initiation. This now encodes a predicted signal sequence much less likely to be used as described previously (23). Similarly, in fibroblast-passaged HCMV strains, deletions interrupt expression of a -chemokine, suggesting that these chemokines may be detrimental or not necessary for in vitro cultivation in some cell types (17, 19). Interestingly, as described earlier, the -chemokine genes in HCMV and HHV-6 are genomic positional homologs, and in HCMV, this locus is associated with endothelial and DC tropism as well as capacity to transfer to lymphocytes, although its receptor specificity has not been characterized (18, 19). HHV-6A does not appear stable in vitro with the U83A chemokine gene, but the potent agonist properties and receptor specificity shown here are consistent with a role in tropism-mediating spread to a variety of hemopoietic cells. Both HHV-6 and HCMV -chemokine genes are hypervariable (23, 43), and possibly these differences may fine-tune interactions with specific cellular populations. For example, the -chemokine homolog in murine CMV has been shown to play a role in monocytic cell-mediated spread of the virus (20, 21). Unlike HHV-6, HCMV also encodes a potent -chemokine, UL146, a neutrophil agonist via interactions with CXCR2 (16) and here also may play a role for neutrophil-mediated spread of the virus.

Of the other viral chemokines described in herpesviruses and poxviruses, HHV-8 vMIPI like HHV-6 U83A also has agonist activity for CCR8, but unlike U83A, it is restricted to this receptor, although similarly potent (44, 45, 46, 47) with an EC₅₀ value for calcium mobilization of specific CCR8-Y3 cells or IL-2-stimulated primary T cells from 0.1 to 1 nM, respectively, and antagonism of the endogenous ligand CCL1 (45). A similarly restricted CCR8-interacting chemokine also was identified in poxviruses but acts only as a potent but selective antagonist, blocking calcium mobilization and chemotaxis of CCR8-HEK293 and CCR8-L1.2 cells, respectively (44, 48, 49, 50). In contrast, HHV-8 vMIPII, acts as a broad antagonist to CCR1, CCR2, CCR3, CCR5, CCR8, CXCR3, CXCR4, XCR1, and CX3CR1, with similar binding affinities for CCR1 and CCR5 of 8 and 5 nM, respectively (49, 51). However, isolated reports using primary eosinophils and Th2 cells suggest some agonist properties (46, 52). HHV-8 vMIPIII has only shown agonist activities of very low potency for CCR4, with chemotaxis of primary Th2 cells at concentrations >100 nM or with specific CCR4-transfected L1.2 cells >500 nM, with no data on calcium mobilization (53). Thus, taken together, all comparisons to date, HHV-6 U83A alone is a novel viral -chemokine showing potent, broad but selective agonist activities to CCR1, CCR4, CCR5, CCR6, and CCR8.

Novel agonist in vaccines or immunotherapeutics for infection and cancer?

From evidence of studies of human and viral chemokine ligands of individual -chemokine receptors, the broad agonism properties of U83A to chemoattract and activate signaling cells bearing CCR1, CCR4, CCR5, CCR6, and CCR8 could act to increase the antigenicity and clearance of tumor cells, often with low immunogenic properties. This also has wide applicability to action of anti-infective vaccines or DNA immunotherapeutic agents. In the case of CCR1 and CCR6, this relates to their expression on immature DCs; thus chemokines that attract and activate these central APCs can induce the antigenicity of copresented vaccine or DNA immunotherapeutic molecule. In studies of tumor or lymphoma model systems, chemokine ligands of CCR1 or CCR6 either injected alone, together with GM-CSF, or fusions with nonimmune tumor Ags, were able to mobilize DC precursors and resulted in increased leukocyte infiltrates, including CD4/8 T cells, PMN, and B cells, with subsequent regression of tumors or lymphomas (54, 55, 56, 57).

Both the human ligands for CCR1 and CCR6, CCL3, and CCL20, also have direct properties in inhibiting myeloid progenitors in colony-formation assays (58). Interestingly, a similar property has been described for a secreted HHV-6 protein (59, 60), with enhanced effects from HHV-6A reported (61, 62). CCR6 ligand, CCL20, also can inhibit the proliferation of chronic myelogenous leukemia progenitors, and thus is a direct biological therapy for cancer with potential application by U83A (63). Furthermore, both CCL20 and -defensins bind CCR6 and have wide antimicrobial activity against a various bacterial and yeast strains (64). Furthermore, both defensins and selected chemokines linked to idiotypic lymphoma Ag gave potent antitumor vaccines (65) with activity as cellular adjuvants or possibly a direct mechanism on the tumor cell membrane. Thus, a molecule such as U83A, which combines the properties of chemokines specific for CCR1 and CCR6, could have combinatorial effects and wider implications for vaccine or immunotherapeutic use in general.

HIV/AIDS immunotherapeutic?

Given that HIV uses chemokine receptors as coreceptors for infection, chemokines and their altered derivatives have been studied as inhibitory factors to virus entry (66, 67). In vivo, HIV primarily uses two receptors, CCR5 and CXCR4, which characterize mainly monocytic and T cell tropic lines, although not exclusively. Further in vitro assays have shown use of other chemokine receptors, including the U83A targets described here, CCR8 and CCR4, but much depends on the relative densities of cell surface expression of these receptors (68, 69, 70, 71). Preliminary results, to be described elsewhere, show blocking by both U83A and U83A-Npep using a CCR5-specific HIV strain (D. Dewin, M. Cleveland, G. Gough, and U. A. Gompels, unpublished observations), consistent with the high-affinity interactions and CCR5-directed chemotaxis inhibition by U83A-Npep demonstrated in this study. Similarly, CCR8-specific agonist vMIPI and broad chemokine antagonist vMIPII have both been shown to inhibit HIV infection in vitro (46, 51).

The distinction for U83A with its broad -chemokine agonist properties is that it has the potential to inhibit HIV through blocking chemokine coreceptors while at the same time acting as a cellular adjuvant to enhance immunogenicity chemoattracting cellular mediators of immunity as described above, thus a novel property combining drug-like inhibitory activity with an immunotherapeutic. Interestingly, results from in vivo studies show that HHV-6 viral loads detected in the blood are lowered coincident with the depletion of its target cell, the CD4 T lymphocyte during HIV/AIDS progression (72, 73, 74). With HHV-6A infection, this depletion in the blood also may enhance HIV replication in other cell types by the removal of a natural, albeit viral, chemokine inhibitor.

Antagonist properties of U83A-Npep and autoimmune disease?

There is evidence for roles of CCR1, CCR4, CCR5, CCR6, and CCR8 in autoimmune inflammatory diseases such as CNS inflammatory disease (MS and Alzheimer’s disease), rheumatoid arthritis, asthma, diabetes, and transplantation rejection, as well as some evidence for other autoimmune disorders such as atherosclerosis, inflammatory bowel disease, and systemic lupus erythematosis. There also could be a role in decreasing tumor infiltrates as described. Thus, there may be therapeutic applications of the antagonist U83A-Npep or possibly other modified versions of U83A in these conditions (67, 75, 76, 77, 78, 79, 80, 81). Such applications also have been described for other viral chemokine antagonists, such as vMIPII or MC148, or modified cellular chemokines. The difference for U83A-Npep or modified U83A is the range of interactions, and for U83A-Npep, its small size, 7 kDa, thus potentially reduced antigenicity over other biologicals, with preliminary results in mice and rabbits consistent with this.

HHV-6 U83A and neuroinflammatory disease

The most interesting area with respect to the biology of HHV-6 is in association with CNS disease, in particular MS. Both CCR1 and CCR5 are main targets in animal models of MS, and experimental autoimmune encephalitis (EAE). A CCR1 antagonist decreases clinical and histopathological disease (82, 83), and in situ hybridization and immunohistochemistry showed CCR1 in early actively demyelinating plaques in monocyte-derived macrophages (84), consistent with data from CCR1 knockout, which showed reduced onset of EAE (85). In addition, in human MS brain sections, both CCR1 and CCR5 were on perivascular monocytes, with CCR1 also on parenchymal monocytic cells (86, 87). Thus, an antagonist of CCR1 or CCR5, such as U83A N-pep made during latent or early infection, could control CCR1- or CCR5-derived involvement in MS, whereas the full-length U83A, made late in lytic primary or reactivated infection, may enhance disease through CCR1 or also CCR8, recently identified in phagocytic macrophages in MS lesions (40). Thus, a relationship with HHV-6A infections may be complex. Studies of HHV-6 and MS are complicated by HHV-6 neurotropism and commensal brain infections, where latent viral DNA of both variants can be detected in normal and MS postmortem brain by PCR (88). However, where active infections are identified by gene expression studies in sera or CSF, recent data show a relationship with MS, with a subset of patients who have either reactivated or primary infections with HHV-6A, (10, 11) and as reviewed (1). In contrast, where comparisons have been made, no MS links have been established with the closely related HHV-7, also neurotropic and a brain commensal (89). These roseoloviruses share almost all of their genes except, strikingly, the U83 chemokine gene, which is deleted in HHV-7 (2, 90). This, combined with the functional properties described in this study for U83A, suggests that, during active infection, U83A may play a role in some MS patients and could be a possible therapeutic target. Furthermore, the CCR1 blocker, U83A-Npep, could be considered a natural immunotherapeutic, and possibly the novel splicing mechanism that regulates its expression (23) also could be considered a unique target for this and other neuroinflammatory disease.

	Acknowledgments

 
We thank Drs. Peter Ertl (GlaxoSmithKline) for help in scaling up bacterial recombinant expression and Maggie Smith (London School of Hygiene and Tropical Medicine) for peptide synthesis. U. A. Gompels also thanks Drs. Eddie Littler and Graham Darby for support at the former GlaxoWellcome and the Biotechnology and Biological Sciences Research Council (U.K.) for ongoing project funding.

	Disclosures

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
In association with the University of London, School of Hygiene and Tropical Medicine, D. R. Dewin and U. A. Gompels have a patent pending on cytokines.

	Footnotes

 
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

¹ This work was supported by a Royal Society Industry Fellowship (to U.A.G.); a Biotechnology and Biological Sciences Research Council/Cooperative Awards in Sciences of the Environment Studentship (to D.R.D. with U.A.G.), undertaken partially at GlaxoSmithKline, Stevenage, United Kingdom; and a Biotechnology and Biological Sciences Research Council Project Grant (to U.A.G., sponsoring J.C.).

² Address correspondence and reprint requests to Dr. Ursula A. Gompels, Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, University of London, Keppel Street, London WC1E 7HT, U.K. E-mail address: ursula.gompels{at}lshtm.ac.uk

³ Abbreviations used in this paper: HHV-6, human herpesvirus 6; HHV-6A, HHV-6 variant A; HHV-6B, HHV-6 variant B; MS, multiple sclerosis; HCMV, human CMV; DC, dendritic cell; RP-HPLC, reversed-phase HPLC; EK, enterokinase; CHO, Chinese hamster ovary; FLIPR, fluorometric imaging plate reader; CI, chemotactic index; RFU, relative fluorescent unit; ¹²⁵I-, ¹²⁵I-labeled.

Received for publication May 25, 2005. Accepted for publication October 26, 2005.

	References

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 

1. Gompels, U. A.. 2004. Roseoloviruses: human herpesviruses 6 and 7. A. J. Zuckerman, and J. E. Banatvala, and J. R. Pattison, and P. D. Griffiths, and B. D. Schoub, eds. Principles and Practice of Clinical Virology 5th Ed.147-168. John Wiley & Sons, Chichester.
2. Gompels, U. A., J. Nicholas, G. Lawrence, M. Jones, B. J. Thomson, M. E. Martin, S. Efstathiou, M. Craxton, H. A. Macaulay. 1995. The DNA sequence of human herpesvirus-6: structure, coding content, and genome evolution. Virology 209: 29-51. [Medline]
3. Isegawa, Y., T. Mukai, K. Nakano, M. Kagawa, J. Chen, Y. Mori, T. Sunagawa, K. Kawanishi, J. Sashihara, A. Hata, et al 1999. Comparison of the complete DNA sequences of human herpesvirus 6 variants A and B. J. Virol. 73: 8053-8063. [Abstract/Free Full Text]
4. Dominguez, G., T. R. Dambaugh, F. R. Stamey, S. Dewhurst, N. Inoue, P. E. Pellett. 1999. Human herpesvirus 6B genome sequence: coding content and comparison with human herpesvirus 6A. J. Virol. 73: 8040-8052. [Abstract/Free Full Text]
5. Billstrom, M. A., G. L. Johnson, N. J. Avdi, G. S. Worthen. 1998. Intracellular signaling by the chemokine receptor US28 during human cytomegalovirus infection. J. Virol. 72: 5535-5544. [Abstract/Free Full Text]
6. Kasolo, F. C., E. Mpabalwani, U. A. Gompels. 1997. Infection with AIDS-related herpesviruses in human immunodeficiency virus-negative infants and endemic childhood Kaposi’s sarcoma in Africa. J. Gen. Virol. 78: 847-855. [Abstract]
7. Akhyani, N., R. Berti, M. B. Brennan, S. S. Soldan, J. M. Eaton, H. F. McFarland, S. Jacobson. 2000. Tissue distribution and variant characterization of human herpesvirus (HHV)-6: increased prevalence of HHV-6A in patients with multiple sclerosis. J. Infect. Dis. 182: 1321-1325. [Medline]
8. Soldan, S. S., T. P. Leist, K. N. Juhng, H. F. McFarland, S. Jacobson. 2000. Increased lymphoproliferative response to human herpesvirus type 6A variant in multiple sclerosis patients. Ann. Neurol. 47: 306-313. [Medline]
9. Kim, J. S., K. S. Lee, J. H. Park, M. Y. Kim, W. S. Shin. 2000. Detection of human herpesvirus 6 variant A in peripheral blood mononuclear cells from multiple sclerosis patients. Eur. Neurol. 43: 170-173. [Medline]
10. Alvarez-Lafuente, R., V. De las Heras, M. Bartolome, J. J. Picazo, R. Arroyo. 2004. Relapsing-remitting multiple sclerosis and human herpesvirus 6 active infection. Arch. Neurol. 61: 1523-1527. [Abstract/Free Full Text]
11. Rotola, A., I. Merlotti, L. Caniatti, E. Caselli, E. Granieri, M. R. Tola, D. Di Luca, E. Cassai. 2004. Human herpesvirus 6 infects the central nervous system of multiple sclerosis patients in the early stages of the disease. Mult. Scler. 10: 348-354. [Abstract/Free Full Text]
12. Hall, C. B., M. T. Caserta, K. C. Schnabel, C. Long, L. G. Epstein, R. A. Insel, S. Dewhurst. 1998. Persistence of human herpesvirus 6 according to site and variant: possible greater neurotropism of variant A. Clin. Infect. Dis. 26: 132-137. [Medline]
13. Byington, C. L., D. M. Zerr, E. W. Taggart, L. Nguy, D. R. Hillyard, K. C. Carroll, L. Corey. 2002. Human herpesvirus 6 infection in febrile infants ninety days of age and younger. Pediatr. Infect. Dis. J. 21: 996-999. [Medline]
14. De Bolle, L., J. Van Loon, E. De Clercq, L. Naesens. 2005. Quantitative analysis of human herpesvirus 6 cell tropism. J. Med. Virol. 75: 76-85. [Medline]
15. Murphy, P. M., M. Baggiolini, I. F. Charo, C. A. Hebert, R. Horuk, K. Matsushima, L. H. Miller, J. J. Oppenheim, C. A. Power. 2000. International Union of Pharmacology. XXII. Nomenclature for chemokine receptors. Pharmacol. Rev. 52: 145-176. [Abstract/Free Full Text]
16. Penfold, M. E., D. J. Dairaghi, G. M. Duke, N. Saederup, E. S. Mocarski, G. W. Kemble, T. J. Schall. 1999. Cytomegalovirus encodes a potent -chemokine. Proc. Natl. Acad. Sci. USA 96: 9839-9844. [Abstract/Free Full Text]
17. Akter, P., C. Cunningham, B. P. McSharry, A. Dolan, C. Addison, D. J. Dargan, A. F. Hassan-Walker, V. C. Emery, P. D. Griffiths, G. W. Wilkinson, A. J. Davison. 2003. Two novel spliced genes in human cytomegalovirus. J. Gen. Virol. 84: 1117-1122. [Abstract/Free Full Text]
18. Gerna, G., E. Percivalle, D. Lilleri, L. Lozza, C. Fornara, G. Hahn, F. Baldanti, M. G. Revello. 2005. Dendritic-cell infection by human cytomegalovirus is restricted to strains carrying functional UL131-128 genes and mediates efficient viral antigen presentation to CD8⁺ T cells. J. Gen. Virol. 86: 275-284. [Abstract/Free Full Text]
19. Hahn, G., M. G. Revel