Hepatitis B virus binds to peripheral blood mononuclear cells via the pre S1 protein

203

Hepatitis B virus binds to peripheral blood mononuclear via the pre Sl protein

P. Pontisso,

G. Morsica.

M.G.

Ruvoletto.

R. Zambello.

C. Colletta,

L. Chemelloand

cells

A. Alberti

hli,U,” <,I,wrdic,,in ClinicI. uniw,r,,ir di fw0”,,. Pdd<,““.,,+ lReCel”Cdz, Fchtuary iv%,)

The hepatitis B virus has been documented in hepatic and extrabepatic compartments, including bane marrow and peripheral blood cells. The viral protein involved in the attachment to human hepatocytes has been identified within the N-tetminus of the pre Sl envelop-e protein. Using recombinant particles containing the pre Sl. pre S2 and S encoded sequences, we studied which virw envelope protein is involved in binding to peripheral blood cells. Mononuclear cells of 20 healthy subjects bound ‘z51-labelled particles, with a Y.Vratio higher than 2.5 (range 2.69-7.77). Binding was abolished by trypsinizalion. B lymphocytes

and monocytes

were found to bind viral particles much more efficiently ,:ompared to T cells and gran-

ulocyter. A mococlonal antibody (MA 18/7). recognizing the (2749) pre Sl sequence, completely inhibited viral particle attachment to PBMC, while anti-pre 52 (Q 19/10) and anti-l (2012) monoclonal antibodies bad no effect. We conclude that the prc Sl sequence is involved in HBV attachment to PBMC and to hepamcytcs. The nature of the cellular attachment site is unknown, but it might be a receptor for physiologic ligands, as occurs with other viruses.

The presence

of hepatitis

B virus (HBV).DNA

in pe-

ripheral blood mononuclear cells (PBMC) has been docomented in patients with chronic hepatitis and with hepatocellular carcinoma

(1). suggesting that these cells are sus-

ceptible to infection by HBV. monocyte marrow,

compartments

Infection of the lymphocyte/

may take place in either

spleen or other lymphoid

blood (l),

tissues or in peripheral

and is most likely initiated

by v?us interaction

with the cell surface. We recently demonstrated pre S encoded

bone

rhat the

domains of the HBV envelope proteins role in HBV attachment to liver plasma

play an important membranes

(23).

the same binding

In this paper we investigated

whether

sites are involved in the interaction

the vims with various types of peripheral

blood cells.

of

Materials and Methads Prepararion

nnd subfracdonorion

Peripheral

blood

were obtained

ofperrpheral

monomiclenr

by Fiwll

cellr

Hypaque

blood cell

(PBAKJ.

PBMC

density-gradient

centri-

fogation from 20 healthy subjects, hho had no evidence of past or present infection by HBV

(serum was negative for

all vtral dntlgen and antibody markers). Granulocyres.

Granulocytes

were prepared

from

the

huffy coat. which formed over the red blood cell pellet fgloiriog

Ficoll

Contaminatma ammonium

Hypaque

density-gradient

centrifugarion.

erythrocytes were lysed by treatment with

chloride

cells were centrifuged

(0.83%.

w/v) for 10 mir.

The white

at 2MNl rpm for 10 min and resos-

204 pended in 0.01 M phosphate (PBS,.

buffer (pH 7.5). 0 15 M NaCl

Mouo~.v:c~. Monocytes were obtained

by plastic adher-

ence. Briefly. PBMC were susocndcd at 10” ccltsiml in complete midium (RPM1 supplemented with 10% FCS) and incubated for h at 37 “C in a Petri dish. After vigorous washing with medium the adherent cells were mechanically removed wilh a plastic scraper. ccnrrifuged at 1000 rpm for 10 min and resuspended in PBS. The final population olwyr contaioed at Ienst YOI positive mono-

I

cytes as datermined by enyrcsaion of CD14 (LeuMJ, Brskton nnd Dikinron. Sunnyvale. CA) on flow cytometry (Cyrutluorograph IIs. Ortho Dia;:nostic System. MA). T nnd B lym,&oryres. T lymphocytes were purified from non-T cells by roztiag with Neuramiaidnse (Sigma Chemicals. St Louis. MO)-treated sheep red blood cells (SRBC) and separation on Ficoll-Hypaque gradient. T cells were recovered by lysis of SRBC with ammonium chloride (U.S3%:. w/v) and resuspended in PBS. The final T cell population contained more then YW. CD2’ (OKTll) and CD3’ (OKT3) (Ottho Diagnostia, Raritan. NJ) lymphocytes, as determined by flow cytometry. Non-resetting cells (B lymphocyte enriched fncrion) were collcctcd at the interface of the Ficoll Hypaque gradient. This nooulation showed mote than 80% CDlY+ (LeulZ) (Beckion aad Dikinnon) lymphocytes cence measurement.

by fluores-

Assay ofviralporfirle hindetg roperipheral blood cells To analyre the binding of wet particles to peripheral blood cells of the 20 healthy subjects, we used radiolabclled recombinant viral particles obtained by transfeclion of yeast cells with a plasmid construction containing the pre St, pre S2 and S regions of the nral genome (4). Immonoblot analysis of these particles showed the presence of two bands at 45 and 39 kDa, corresponding to the presence of both glycosylated and non-glycosylated polypeptides (4). In addition. they were shown to express all three protein domains of the HBV envelope, i.e., the pre Sl, pre S2 and S sequences, when reacted with specific monocloaal antibodies (4). The recombinant particles were radtolabelled with “‘1 using the lactoproridaselglucose oxidase method (Bio-Rad) and free radioactivity was

alive comrol. tubes without cells were incubated with the labelled P;xticles and washed as above. A ratio between cpm of fhc sample and mean cpm of five negative control tubes (SIN) of higher than 2.5 was considered as positive. In order to exclude unspecific binding in our expenmental conditions. ‘*il-labellcd human serum albumin (Sarin Biomedica, Saluggia. Italy) wus used as control Iig:,nd.

In preliminary

experiments

we demonstrated

that re-

combinant particles containing the S gene product (Sift of Genentech Laboratories) or the pre S2 and S derived protein (gift of M.I.. Michel. Pasteur Institut, Paris) could not bind to Bunyan peripheral blood citts. Only recombi. nant particles carrying the pre Sl. pre S2 and S encoded proteins, the same as those used in the present study, were capable

characterise

the

pattern of interaction between these recombinant cles and PBMC. increasing numbers of (5~10i-10~10b) were added to a fixed number of ‘IsI cles (10s cpm in 50 /II). incubated for 45 min at washed three times and counted after pelleting.

of binding

particells, parti4 ‘C,

The time-dependence

(5). To better

of the binding reaction was eval-

uated by incubating labelled particles with 5.136 cells for 15 s, and 5. lo,30 and 60 min in ice. After three washes with PBS-CFcellspellets were counted. To define the exact specificity of the binding site on recombinant particles, inhibition experiments were performed using monoclonal antibodies (Mabs). The following monoclonal antibodies, which reacted with the above particles (data not shown), were used: Mab MA 1617, which recognises XI epitope between position 27-49 of the pre Sl encoded protein; Mab QlY/lO, which is directed against the glycosylated portion of the pre S2 encoded protein (AA 120-130): and the Mab 2012 which recognises the S encoded product of the viral genotne (3). Serial dilutions of these three monoclonals (from 100 to 0.1 &ml in PBS-DF) were mixed I:1 (v/v) with labelled particles (l@ cpm in lOOpI). Following incubation with 5.106 PMBC for 45 mitt at 4 “C, samples washing. To test the binding of recombinant

were counted

after

particles to different

removed using a P6DG desalting gel column. Five million cells in IcOfll of PBS Dulbecco’ Formula (PBS-DF), containing 0.6 mM Cat&, 0.5 mM MgC12, were incubated at 4 “C for 45 min in 5 ml plastic tubes, with 5Opl of labelled pzrticles (in ttiplicst;). weshed by centrif~gation and counted io a gamma counter. Low temperature incuba-

subpopulations of white cells, cells from four different subjects were prepared (Table 1) and the cell count ndjusted to 5106 per sample to be incubated with ‘*$labelled particles for 45 min at 4 “C. Cells were then washed

tion was chosen in order to limit the reaction to the attach-

inp PBMC with PBS containing 25% trypsin (Ciibco BRL) for20 min at 37 “C. Cells were washed three times in PBS

ment phase, since at highertemperawres (i.e., 20-30 ‘C) internalisation of the ligaod cannot be excluded. As a neg

three times and the final pellet was counted. The effect of trypsinisation on the cell binding was studted by incuhat-

and incubated

with labelled

particles.

As a positive con-

cles with a SIN ratio higher than 2.5 (mean net cpm k SD.. 546.7

4213 * 2259.8; mean negative control cpm F SD.. k

152.1).

No binding

was observed

achteved uwth mcrearingamour.rs

when radio-

Mahi

labelled human albumie was substituted for ‘zil-recomblnant putticks. the different

The cwfficient preparations

ranging from 3 to 10%. pared from different binding

activity,

of variation

c~nccntrat,on

On the other hand, PBMC

with the mean (triplicate

m

experiments)

by trypsinisation

B

as shown in the mpre

sentative example of Fig. 1. Time-dependence mcnts revealed high avidity of the cellular since a plateau was obtained

experi-

binding

site,

within the first 10 min of in-

cubation (Fig. 2~.

ofthe virw otrnchmentsite

complete

Table L shows the binding of radiolabellcd particles IF the different

There was a linear relation

fixed amount of labelled particles,

ent Mabs,

used (Fig. 3).

subpopulations

from low healthy HBV-negJti,vesuhjects.

abolished

site on PRMC.

recombinant

whde

prc-

subjects had a discrete variation

between bound cpm and the number of cells used wth

When

iW.

observed with

treatment of the cells. Talus indicating a proteic nature of

finespeci@iry

of Mab MA

and 20/Z did not affect the bmding a: any

from the fame donor ~8s low.

PIN ratio from 2.68 to 7.77. Binding was completely the attachment

OlSilO

for PBMC

particles were reacted with diffcrinhibitton

of binding

fo PBMC

was

recramhmimt

of cells prepared In all fout sub-

jccrr bmding to B lymphocytes and to monocytes was sig-

binding of virus particles to B cells and monocytes via the

nifamrly

pre St protetn domain

higher rhan bmding observed with T lympho-

be related to particle

cytes and with grnnulncytes.

may initiate

cell infection

The nature of the cellular site of attachment virus binding Discussion

or may

uptake by antigen-presenting

to peripheral

blood

cells.

invalved IV

cells is unknown.

It

might he a cell receptor used by the %iris, which mimics the physiologic ligand. This occurs with other viruses (10)

The hepatitis I3 virus, contrary to common bebef, is not stricrly hcpatotropr. tablishcd

that tissues from

bone marrow. may contain HBV-DNA

the kidney,

as well as peripheral Hl3V-DNA

blood

The

skin,

detection

of

cells has

interest in relation to possible effects on

immune system. Published studies do nol seem to of HBV-DNA

in a rpeclfic subpopulation

of PBMC,

reported

positivity withalthough

the presence of HBV-DNA

such as adenoviruscs. leukemia

Epstein

Barr virus, human T cell

virus, human immunodeficiency

virus and rabies \‘irus. Computer-assisted

virilr. vaceini’d protein similari-

w research of the putative pre SI sequence of the virus in-

cells (l&7)

mononuclear

indicate compartmentation

al. (8)

pancreas,

blood

sequexes

in peripheral

geucrared goxt the hat

Indeed, several studies hove now es-

Yoffe

et

mainly

in

volvl?

in the binding has indicated

constant domain of humw

high homology with a

IgA (Ref.

and ?&irk 13, personal communicaiion).

9, and Howard

CR

Receptors for IgA

are present on several human cells, including hepatoeytes (11).

monacytes

dttd polymorphonuclear

cells (12)

and

might indeed represent the cellular receptors for HBV.

manocytes and B cells. In support of these observations, WC found that recombinant

particles bind preferentially

to

Acknowledgements

mnnocytes and B lymphocytes and, to a lesset extent. to T lymphocyte

and

granulocytes.

tarhmrnt

could

be

The authors are deeply grateful to Professor W.H. lich (Hygiene

the same site which is involved in virus at-

for providing

to hepatocytes (3,s).

was indeed inhibited no effcn ws

binding

sequence of the pre Sl pro-

shown to involve the (27-49) tein domain,

This

by anti-pre

The virus-cell tnteraction Sl Mab MA

18.7, while

seen wtth anti-pre S2 and anti-S Mabs. Tbc

Institut,

University

the monaclotml

of Gottingen, antibodies

MA

19110 and 2012, and to R.W.

Ellis and P. Kniskern

Laboratories,

U.S.A.)

of recombinant

Philadelphia, particles.

Ger-

F.R.G.) 1817, Q (Mle:ck

for the generous gift