antibody patterns following interferon therapy in acute and chronic hepatitis B

antibody patterns following interferon therapy in acute and chronic hepatitis B

Journal of Hepatology 1994; 20:47-56 Printed in Denmark. All rights reserved Munksgaard Copenhagen Copyright © Journalt~ Hepatolog.i,1994 Journal of ...

1MB Sizes 1 Downloads 20 Views

Journal of Hepatology 1994; 20:47-56 Printed in Denmark. All rights reserved Munksgaard Copenhagen

Copyright © Journalt~ Hepatolog.i,1994 Journal of Hepatology ISSN 0168-8278

PreS1 antigen/antibody patterns following interferon therapy in acute and chronic hepatitis B M a r i e - A n n e Petit ~, Fabien Zoulim 2, Pascale Berthillon 2, Francis Capel ~, Jisu Li 2, Charles D a u g u e t 3, Carlo Ferrari 4 and Christian Tr6po 2 I I N S E R M U131, Clamart; -'INSERM U271, Lyon; JViral Oncology Unit, Pasteur Institute, Paris, France," and 4Cattedra Malattie Infettive, Universit~ di Parma, Italy

(Received 27 April 1992)

The relation between preS1 antigen/antibody system and different phases of hepatitis B virus infection were studied in 425 serum samples from 50 hepatitis B patients before, during and after antiviral therapy using interferon alone or in combination with corticosteroid withdrawal. A typical profile of self-limited acute hepatitis B was characterized by hepatitis B virus-DNA clearance using polymerase chain reaction and preS antigens using monoclonal radioimmunoassays and by antibody responses to the middle and the large HBs proteins (gp33/gp36 and p39/gp42) using immunoblotting quantitative analysis. After interferon therapy in patients with protracted hepatitis B, complete eradication of the virus was observed in 70% of patients, and antibody response directed to middle HBs and large HBs proteins could be induced. Conversely, this antibody response was never detected in follow-up studies of chronic active hepatitis B patients who responded well to antiviral therapy and lost HBs, preS2 and preSl antigens. Most interesting, in 50% of patients with HBeAg-positive chronic active hepatitis B who received combination therapy and in 67% of patients with anti-HBepositive chronic active hepatitis B given interferon alone, the elevated serum preS1Ag/HBsAg ratio persisted after treatment was discontinued and even increased until the end of follow-up when hepatitis B virus DNA was undetectable in serum by the conventional hybridization technique. This rebound of preS1 antigen expression following antiviral therapy in patients with chronic active hepatitis B may indicate virus persistence, suggesting the possibility of relapse through wild-type hepatitis B virus or the emergence of hepatitis B virus mutants. © Journal of Hepatology. Key words: Hepatitis B infections; Interferon therapy; Mutants; PreS1 antigen/antibody

The hepatitis B virus (HBV) envelope contains polypeptide sequences coded by the preS regions of the viral genome expressing two additional preSl and preS2 antigenic specificities which are distinct from HBsAg (1). Both preS2 and preSl domains have been shown to contain host receptor binding sites (2,3) and to be involved in the attachment of HBV to hepatocyte receptors. HBV can bind indirectly via polymerized human serum albumin (pHSA) as an intermediate molecule through the preS2 domain (2,4), or directly to the host-cell membrane through the preS1 domain (3-5). Monoclonal antibody (MAb) F35.25 has been shown to have excellent recog-

nition of the region 21--47 of the preSl sequence, which corresponds to the hepatocyte-receptor binding site on HBV (3), independently of d/y changes (6,7). Therefore, we developed a double-radioimmunoassay (RIA) to assess accurately the many preSl-epitopes expressed on the HBV/HBsAg particles in serum samples from HBV-infected individuals (8). Using this method, the expression of preSl antigen (preSlAg) correlated well with the level of HBV replication in patients with chronic active hepatitis B, especially among anti-HBe carriers (8). The aim of the present study was to investigate the clinical significance of the preS1 antigen/antibody system in

Correspondence to." Marie-Anne Petit, Ph.D., INSERM, Unit6 131, Immunopathologie & Immunologie Virale, 32 rue des Carnets, 92140 Clamart, France.

48 serum of patients at different stages of HBV infection: protracted hepatitis B, HBeAg- or anti-HBe-positive chronic active hepatitis B (CAH-B) associated or not with cirrhosis, following treatment with interferon (IFN) according to different protocols (9-11), including combination with corticosteroid withdrawal (10). PreS antigens (preS2Ag and preS lAg) were determined before, during and following antiviral therapy and compared with classical serum HBV markers of viral replication (hepatitis B e antigen and antibodies, HBV-DNA and DNA polymerase) and serum aminotransferase levels (AST and ALT). Follow-up studies of patients with selflimited acute hepatitis B (AH-B) were included to assess qualitative and quantitative analysis of the antibody response to the middle (M) and large (L) HBs proteins (gp33/gp36 and p39/gp42) by the immunoblot technique, and thus to provide a typical serologic profile of patients who recover spontaneously with complete eradication of virus_ Results of this study showed that IFN therapy was more effective in patients with protracted hepatitis B than in patients with HBeAg- or anti-HBe-positive CAH-B. In addition, a rise in preS1Ag expression in serum of patients with CAH-B considered good responders to antiviral therapy (loss of HBeAg and HBV DNA in serum) may indicate viral persistence, which may or may not be associated with the emergence of and gradual takeover by defective mutated HBV strains.

Patients and Methods

M.-A. PETIT et al. CAH-B (with histological cirrhosis in 9 of 12) were antiHBe-positive. Only four of 12 (33%) were positive for HBV DNA by hybridization. However, liver biopsy specimens from 11 cases tested for HBcAg by immunofluorescence (IF) were found positive (with cytoplasmic rather than nuclear localization), indicating low-grade HBV replication before IFN therapy. All patients were positive for HBsAg and anti-HBc in serum before the beginning of the study, and negative for antibodies to hepatitis D virus (HDV), hepatitis C virus (HCV) and human immunodeficiency virus (HIV)_ Serial serum samples were available at 1- to 4-week intervals in all 50 patients recruited from 1988 to 1991. Treatment

Only recombinant human a-interferon-2b (rlFNa-2b) was used. Patients with protracted hepatitis B received either placebo or IFN at a dose of 5 million units (MU) subcutaneously (s.c.) 3 times a week (t.i.w.) for 24 weeks (9). HBe-positive patients with CAH-B were randomly assigned to one of the three following treatment regimens: 12 weeks of IFN alone at a dose of 4.5 MU 3 times per week (11); IFN at the same dose but associated with prednisone in decreasing daily doses of 60 mg, 40 mg, and 20 mg, each for 2 weeks, either concomitantly for 20 weeks (10), or sequentially following a 2-week rest. Anti-HBepositive patients with CAH-B received different dose schedules of IFN: 1 MU 3 times per week for 16 weeks or decreasing doses of 8 MU, 5 MU, 3 MU, and 1 MU, every 16 weeks for a year.

Patients

Follow-up studies were performed in 425 serum samples from 50 patients. Seventy serum samples were obtained from ten patients with a natural course of selflimited AH-B. These samples had been analyzed previously for the proliferative response of peripheral blood lympho-mononuclear cells (PBMC) to HBV envelope, core (HBcAg) and e (HBeAg) antigens (12). Ninety-nine serum samples were available from ten patients (four in the placebo group and six in the treated group) with protracted (> 10 weeks but <6 months) hepatitis B infection who were participating in a clinical trial for interferon therapeutic efficacy (9). Two hundred and fifty-six serum samples came from 30 patients with CAH-B diagnosed clinical, biochemical and histological criteria (13). All chronic hepatitis B patients had elevated ALT levels, showing ongoing liver disease activity. Eighteen patients were positive for HBeAg, HBV DNA (detected by hybridization) and DNAp, indicating active viral replication before antiviral therapy. Six of these received IFN alone, and 12 concomitant (8/12) or sequential (4/12) prednisone withdrawal and IFN_ The remaining 12 patients with

Classical HB V markers

HBsAg, anti-HBs, anti-HBc, HBeAg and anti-HBe were determined with commercially available RIA kits (Abbott Laboratories, North Chicago, IL). DNAp activity was assayed as previously described (8,14), and considered positive when the change in counts per minute exceeded 50 cpm. Serum HBV-DNA was determined by the spot-hybridization technique (15). Liver biopsies were performed before antiviral treatment in patients with elevated ALT levels. The immunoflorescence (IF) method used to detect HBcAg in the liver has been described previously (16). PCR assays

HBV-DNA detection was monitored by the Polymerase Chain Reaction (PCR) assay as described in detail elsewhere (9,17). A rapid oligonucleotide detection method with a stop codon in the distal preC region (18) was performed to screen for common HBV variants in selected serum samples from patients with suspected mutated HBV-DNA forms.

PreSIAg/Ab AND RESPONSE TO INTERFERON

Detection of preS2 and preS1 antigens by RIAs The quantitative expression of preS2Ag and preS1Ag in serum was studied using well-defined monoclonal antibodies (MAbs) (6,7,19,20) in polyclonal-monoclonal RIAs, previously described in detail (8). Briefly, polystyrene beads were coated with rabbit polyclonal IgGs to HBsAg, and incubated with serial 10-fold dilutions of human sera. Then, murine MAbs to HBsAg (F39.20), preS2Ag (F124), or preS1Ag (F35.25) were added, and ~25Ilabeled F(ab')2 fragment of anti-mouse immunoglobulins (Ig) (Amersham France) was used as tracer_ All serum samples were tested at three dilutions (10 -~, 10 -2 and 10-3). Results were expressed as cpm of ~25I-labeled antibody bound. Since HBsAg, preS2Ag and preS1Ag assays exhibit the same sensitivity, preS2Ag/HBsAg and preS1Ag/HBsAg ratios (as a ratio of c p m × 100) could be calculated throughout follow-up from serum samples diluted 10× from each patient for assessment of the variability of relative proportions of preS2- and preSl-epitopes expressed on HBsAg-positive particles (8).

49 for HBsAg, preS2Ag and preSlAg by means of RIAs described above.

Electron microscopy Fractions of the sucrose gradient were observed directly after negative staining with 0.5% uranyl acetate under a 100 CX electron microscope (JEOL Ltd., Tokyo, Japan).

Results

1. Antibody responses to MHBs (gp33/gp36) and LHBs (p39/gp42) proteins durhzg recovery from acute hepatitis B (AH-B) Nine of the ten patients with AH-B completely cleared the three HBV surface antigens (HBsAg, preS2Ag and preSlAg) and seroconverted to anti-HBe. Four of them (44%) also seroconverted to anti-HBs in less than 6 months. Fig. 1 shows the typical antibody response to MHBs and LHBs proteins in a patient who completely recovered from AH-B, in relation to other viral markers in serum (IgM anti-HBc, anti-HBe and anti-HBs), ALT

Detection of antibodies to MHBs (gp33/gp36) and LHBs (p39/gp42) proteins The western immunoblotting assay (WIBA) was used to identify the polypeptide specificities of preS (preS2 and preSl) antibodies in the sera of patients with spontaneous recovery or recovery following antiviral therapy. Complete HBV particles (subtype ay) purified from serum of an HBeAg-positive chronic carrier were used as an antigenic probe, as previously described (21). Purified viral preparations were denatured by 2% (w/v) SDS and 5% (v/ v) 2-mercaptoethanol (2-ME), and then subjected to electrophoresis in 12_5 % polyacrylamide gels (SDS-PAGE) followed by blotting analysis. Transfer blots were incubated with serum samples diluted 1:20 in blocking buffer (0.01 M Tris HCI, 0.14 M NaCI, pH 7.2 containing 5% bovine serum albumin and 50% normal rabbit serum). Antibody binding was detected with L'5I-labeled F(ab)'2 fragment of anti-human immunoglobulins (Amersham France). To test the specificity of anti-preS1 detection, serum samples were pre-adsorbed with CHO cell line HBsAg (22). To quantify the antibody response, the scanning of blots was performed using an UltroScan XL Laser densitometer (Pharmacia LKB Biotechnology). Results are expressed as area (AU×mm).

Sucrose gradient analysis of viral particles Serum samples (0.2 to 0.5 ml) from one patient with HBeAg-positive CAH-B before and following antiviral therapy were loaded onto a 20 to 60 % linear sucrose gradient in T N buffer and centrifuged for 18 h at 200000×g and 4°C. Fractions were collected and probed

Anli-HBc(IgM)

,s]

Anli-HBe I

Anti-HBs

Q

too

410' 2!

<3

g 5,

I

¢ m

.& 4[ O, 4

8

12

16

20

24

wookg

<1.x z Be HB¢

_=-

Fig. I. Kinetics of antibody responses to preS-proteins in a patient who completely recovered from AH-B, in relation to anti-HBs response (©), ALT levels (O), specific T-cell response and other markers in serum (IgM anti-HBc, anti-HBe). Symbols: total antipreS2 proteins (gp33+gp36) response (A), response to gp33 preS2proteins alone (A), total anti-preSl proteins (p39 + gp42) response (i), response to gp42 preSl-proteins alone (I-I).Significance of stimulation index: +, 2.5-20; ++, 20-50; +++, >50.

50

M.-A_ PETIT et al.

TABLE 1 PreS status (preS2Ag/HBsAg and preSIAg/HBsAg ratios) in patients with protracted hepatitis B before and following interferon therapy Patient no. 1h 2 3 4 5 6

Post-treatment final follow-up

Pre-treatment baseline PreS2Ag/HBsAg ratio"

PreSIAg/HBsAg ratio"

PreS2Ag/HBsAg ratio"

PreS I Ag/HBsAg ratio"

>50% (+ + +) 17% (+) < 10% (+) 50% ( + + + ) 15% (+) 15% (+)

< 10% (+) 33% (+ +) 14% (+) < 10% (+) 24% (++) 22.5% (++)

negative negative negative negative preS2+lim. <10% (+)

negative negative negative negative 31.5% (++) 25% (++)

Ratio ofcpm× 100 (%) at a 1:10 dilution for preS-positive serum samples. b Patient no_ I corresponds to patient B in Fig. 2b_

levels and specific T-cell responses to nucleocapsid antigens (HBcAg and HBeAg). The maximal response to MHBs proteins (gp33+gp36) was associated with a transient but marked exacerbation of hepatitis, as indicated by a sudden rise in A L T levels. This coincided with the disappearance o f l g M anti-HBc and the appearance of anti-HBs in serum, and preceded a vigorous temporary T cell response to nucleocapsid antigens (HBcAg and HBeAg). Total anti-preS2 response declined but persisted for a longer period together with high anti-HBs production. Maximum response to LHBs proteins (p39+gp42) appeared slightly delayed. Total anti-preSl response gradually decreased and became negative within 6 months. Interestingly, the humeral responses to monoglycosylated MHBs protein gp33 and glycosylated LHBs protein gp42 were found to be higher (gp33/gp33+gp36 mean ratio= 72% and gp42/p39+gp42 mean ratio=66%, respectively, Fig. 1) than responses to diglycosylated MHBs protein gp36 and unglycosylated LHBs protein p39.

II. Response to IFN therapy in patients with protracted hepatitis B Ten patients were included in this study: four in the placebo group and six in the treated group. All patients showed persistent HBeAg and HBV D N A in serum as detected by dot-blot hybridization 2 to 3 months after the onset of symptoms and failed to normalize serum A L T levels. The preS status (preS2Ag/HBsAg and preS1Ag/HBsAg ratios) in serum of the six treated patients before and following I F N therapy are given in Table 1. The pre-treatment baseline preS1Ag/HBsAg ratio was low ( < 15%) in three of the four responders (patients no. 1 to 4) and higher (24% and 22.5 %) in the two non-responders (patients no. 5 and 6) (Table 1). Fig. 2a shows the serological profile of patient A. This patient spontaneously normalized A L T levels, cleared the three HBV surface antigens and H B V - D N A from serum, and seroconverted to both anti-HBe and anti-HBs after receiving placebo. To our surprise, no antibody response to

MHBs (gp33/gp36) and LHBs (p39/gp42) proteins could be detected by western immunoblotting assay (WlBA) throughout follow-up (Fig. 2a). Conversely, there was a strong signal in W I B A for antibody response to gp33/ gp36 and p39/gp42 proteins in one patient (B) treated with I F N who responded well and exhibited seroconver:1

Trealmant: PlaCebo antigens

14

~--N 10

~

preSlAQpreS2Ag 500

•1"

100 ~

o:

[]

10

20

MONTHS HBV-ONA (PCR)

+** ** *-~ ÷*

antI-HBg ( R I A )

negative

snll-preS (WIBA) anti-pre$2 anti-preS1

negatLve negacLve

negat;Lve negot;|ve

b.

Trealmenl: InUonA HBV surface anllgens Z

12'

t3OO ~

lo.

i,[

=..

,F~..~_

~reS2~g ~reSlAO

"

8'

200

-

6.

°"2. 0

"~--

2 HBV-DNA (PCR)

anII-HBs (RIA) ann-preS (WIBA) anll-preS2 anti-preS1

~

4

+** + ÷ . . ÷ . .

. . . . . -

6 8 MONTHS

,

"~

10

12

14

~ ÷ + .

÷

*

+ // ~,*-*

+ . ++÷ ~ . : +4- ~ ++ :

negative

Fig. 2. Clinical course of two patients with protracted hepatitis B. a, Patient A given placebo who spontaneously seroconverted to antiHBe and anti-HBs with complete elimination of the virus, b, Patient B given interferon who responded well and exhibited a typical sereconversion illness. Antibody response to both preS2 and preSl proteins was detected and quantified by western immunoblotting assay (WIBA). PCR=polymerase chain reaction. RIA=radioimmunoassay.

PreSlAg/Ab AND RESPONSE TO INTERFERON

51

TABLE 2 PreS status (preS2Ag/HBsAg and preS|Ag/HBsAg ratios) in patients with HBeAg-positive chronic active hepatitis-B before and following interferon (IFN) therapy given alone Patient no.

1h 2 3h 4 5c 6

Pre-treatment baseline

Post-treatment final follow-up

PreS2Ag/HBsAg ratio~

PreS1Ag/HBsAgratio"

PreS2Ag/HBsAg ratio"

PreSIAg/HBsAgratio~'

12"/'o (+) <10% (+) 45% ( + + ) >50%(+++) < 10% (+) 16% (+)

12% (+) 19.5% (+) 12.5% (+) 15% (+) 24% (+) 24'/0 (+)

15% (+) <10% (+) 42% ( + + ) 50%(+++) negative <10% (+)

20% (+) 18% (+) 13% (+) 11%(+) negative 21.3% (+)

:' Ratio o f c p m × 100 (%) at a I:10 dilution for preS-positive serum samples. b Persistence of HBeAg and HBV DNA after IFN therapy (NR). c Super-responder (R+).

sion illness with an ALT flare (Fig. 2b). This response was previously described in a patient who recovered from AHB (cf. Fig. 1).

IlL Persistent or hwreased preSIAg expression following antiviral therapy in patients with CAH-B Until now, the loss of HBeAg and HBV D N A from serum tested by dot-blot hybridization has been found to reflect an inhibition of HBV replication and be a strong indicator of response to antiviral therapy. However, most patients who seroconvert to anti-HBe are still HBsAg positive after treatment_ At present, the serum preSIAg determination can be considered an accurate predictor of virus clearance (8). Therefore, we defined: 1) a super-response ( R + ) as the sustained clearance of both HBsAg and preS1Ag in association with elimination of HBV markers of viral replication (HBeAg and HBV DNA), corresponding to complete eradication of the virus; and 2) a partial response (PR) as the loss of HBV D N A with persistent HBsAg associated or not with an elevated serum preSlAg/HBsAg ratio at final follow-up. Patients with HBeAg-positive CAH-B. Of 18 patients with HBeAg-positive CAH-B, six received IFN alone and 12 IFN in combination with corticosteroid withdrawal. After 10 months of I F N therapy alone, one patient (no. 5, Table 2) dropped his HBsAg titer and became negative for both preS2Ag and preS1Ag (R+). Two patients (no. 1 and 3, Table 2) did not respond and retained persistent HBV DNA, HBeAg and preSIAg. The three other patients (no. 2, 4 and 6, Table 2) cleared serum HBV DNA as detected by dot-blot hybridization, but were still HBsAg and preSIAg positive (preSlAg/HBsAg mean ratio =16.8%) after IFN therapy (PR), suggesting that they remained slightly viremic as confirmed by positive results for HBV D N A by PCR (17)_ In the 12 patients receiving IFN in combination with corticosteroid withdrawal (concomitantly or sequentially),

three of 12 did not respond, while three others were superresponders (R+) showing full recovery with complete clearance of the virus (Fig. 3)_ HBV DNA, HBeAg, HBsAg, preS2Ag and preSlAg were eliminated sequentially, followed by seroconversion to both anti-HBe and anti-HBs 1 year after treatment. Surprisingly, no antibody response to MHBs and LHBs proteins could be detected by WlBA throughout follow-up (Fig. 3). The remaining six (50°/,,) were partial responders (PR), and showed typical

Treatment:

prednlsoloneI ] l l l l I I I I I I I I I 1 ! Inlron A

IOOO

20

10

soo .-I

I HBV DNA

(DoI-Blol)

HBeAg/AnlI-HBe anli-Hna

anll-preS

(RIA)

(WIBA)

~

#

2 4r~+

+

34 6.7 MONTHS ~I~

I~_°bbb~b*~*~ nega tlve

negative

813

negative

-

I -

[]

negative

Fig. 3. Typical profile of a super-response (R+) in an HBeAg-posi-

tive chronically-infected patient receiving interferon in combination with corticosteroid withdrawal. Complete elimination of HBsAg (O), preS2Ag (A) and preSIAg (I) was associated with resolution of hepatitis (ALT normalization), and seroconversion to both antiHBe ( [], HBeAg-positive period, I , anti-HBe-positive period) and anti-HBs. No antibody response to preS proteins was detected by western immunoblotting assay (W1BA) throughout follow-up. RIA = radioimmunoassay.

52

M.-A. PETIT et al_

serologic preSIAg profiles (Fig.4a). The pre-treatment baseline preSIAg/HBsAg ratio (=25%) turned to zero after

preS1Ag/HBsAg ratio rose again markedly (up to 30 to 40%) during post-treatment follow-up, first in association

1 m o n t h of therapy when D N A p decreased and HBV D N A disappeared from serum (Fig. 4a). Nevertheless, the

with the persistence of HBeAg and then despite seroconversion to anti-HBe 1 year later (Fig. 4a).

b.

15L-I ( pre-TRT

/

a, Triilmlnl:

Pmd~lzolone InlrOn A

baseline

88)

JJ (post-TRT follow-up, Sept 89/Apr 90)

m11

pOII-TRT

pra-TRT

108

May

follow-up

baseline_ /

-

1OO

1G

io

/

[ o

]0

lOO

::)

20 o 0

!

Jan 1988

,

1

Jul

Jan 1989

Jul

Jan It990

o

X

10

o<

E

u HBV DNA (Dot-Blot)

+ ÷ + ++ * ~+ ÷ -

HBoAg/AnII.HBe

neqatlve

fraction number

C.

"~.

It,,

~ ~

Fig. 4. Typical profile of a partial response (PR) in an HBeAg-positive chronically-infected patient receiving interferon in combination with corticosteroid withdrawal, a, Evolution of the preSIAg expression (serum preSIAg/HBsAg ratios) in relation to ALT, DNAp and HBV DNA levels, and the HBeAg ( [] )/anti-HBe ( • ) system. b, Isopycnic sucrose gradient analysis of the first serum sample from May 1988 (I, pre-TRT baseline) and of the serum samples from Sept. 1989 to Apr. 1990, 1 to 2 years after beginning the treatment (II, post-TRT follow-up). Fractions were probed for HBsAg (O), preS2Ag (&) and preS1Ag (11) by means of PAb-MAb radioimmunoassays. The sucrose concentration (O) was assayed by refraction, c, Fraction 10 from b, II was further subjected to electron-microscopic examination.

PreS1Ag/Ab AND RESPONSE TO INTERFERON To investigate whether the preSl-epitopes were expressed on complete or defective viral particles (VPs), serum samples obtained in 1988 (pretreatment baseline) and in 1989-1990 (post-treatment follow-up) were subjected to isopycnic sucrose gradient centrifugation. In the serum sample from 1988 (I, Fig. 4b), preS1Ag was recovered in fractions 11 and 12 banding at 40% sucrose, suggesting the presence of preSl-epitopes on complete virions (8) during the pretreatment period. In serum samples from 1989-1990 (II, Fig. 4b), preS1Ag was recovered in fractions 9 to 11 banding at 30% sucrose. Examination of the last fraction 10 (II Fig. 4b) by electron microscope revealed the presence of empty and full modified spherical VPs with a reduced diameter o f 30 to 35-nm (Fig. 4c). Filamentous forms and many degraded protein structures could be also observed (Fig. 4c). Patients with anti-HBe-positive CAH-B. Using an oligonucleotide detection method, one-point preC mutation was detected in four out o f the six (67%) patients tested by PCR (Table 3), indicating the presence of HBeAg minus mutants before I F N therapy. Table 3 shows that the pre-treatment baseline preSIAg/HBsAg ratio in the 12 patients with antiHBe-positive C A H - B exceeded 10% (mean=15.3%) in 8 out of 12 (67%), was <10% in 3 (no. 8, 10 and 11) and null in only one (no. 4). After 8 to 16 months of I F N therapy, the data showed that serum HBV D N A was undetectable by dot-blot hybridization in 100% of patients, and normal levels of serum A L T were achieved in 50% among them. In two patients initially positive for HBV D N A , A L T levels remained abnormal throughout the observation period. One patient (no. 1, Table 3) completely cleared HBsAg,

53 preS2Ag and preS1Ag ( R + ) after a 2-year follow-up evaluation. Four patients (no. 3, 6, 7 and 9, Table 3), one of whom (no. 7) had very high baseline preS1Ag expression (preSIAg/HBsAg=30%), showed a decrease in preS1Ag/HBsAg ratio during I F N therapy (Table 3). Conversely, the majority of patients (7/12, approximately 60%) exhibited an increase in preSIAg expression (preS1Ag/HBsAg mean ratio=22.6%) following I F N therapy (Table 3). Four patients (no. 4, 8, 10 and 11, Table 3) with low (< 10%) or null pre-treatment baseline preSl Ag/ HBsAg ratio experienced a strong rebound of preSIAg/ HBsAg ratio (mean=25%) following treatment with IFN_

Discussion

In the present study, the clinical significance of preS l A g expression and antibody responses to MHBs and LHBs proteins were investigated in patients with selflimited acute hepatitis B, protracted hepatitis B or chronic active hepatitis B in response to antiviral therapy. It is known that the complete elimination of infectious virus is usually associated with the sustained disappearance of HBV D N A and seroconversion to anti-HBe and antiHBs, accompanied by a normalization of serum alanine aminotransferase levels and histological improvement. In a recent study (8), we showed that the detection and quantification of preS1Ag in human serum provided a new, clinically relevant diagnostic assay to assess HBV replication in patients with CAH-B, especially among anti-HBe carriers. Therefore, it could be expected that preS1Ag testing might also provide a convenient means

TABLE 3 PreS status (preS2Ag/HBsAg and preSlAg/HBsAg ratios) in patients with anti-HBe-positive CAH-B before and following IFN therapy Patient Pre-treatment baseline

Post-treatment final follow-up

110.

Ib 2¢ 3 4' 5~ 6 7 8' 9 10~ 11~ 12¢

PreC mutation detection method (ref. 18) Oligoprobe

PreS2Ag/HBsAg PreSIAg/HBsAg ratio u ratio •

PreS2Ag/HBsAg PreS1Ag/HBsAg ratio ~ ratio ~

M0 (nonmutated)

MI (one pointmutated)

M2 (two pointmutated)

19.5% (+) 41.5%(++) 36% (++) preS2+lim, 33% (++) 39.3% (+ +) >50% (+++) >50% (+++) 28.5% (+) >50% ( + + + ) 17% (+) >50% ( + + + )

negative 36%(++) 35% (++) 37.5% (++) 30% (++) 36.4% (+ +) >50% (+++) 37.9 (++) 35% (++) 34.8% (++) 10% (+) >50% (+++)

n.d_ n.d. + n.d. n.d_ + + n.d_ n.d. -

n.d. n.d. n_d. n.d. + + + n.d. n.d. +

n.d. n.d. n.d. n.d. + + n.d. n.d. +

17.7% (+) 11%(+) 13% (+) negative 12% (+) 12.8% (+) 30% (++) <10% (+) 11.2% (+) preSl +lim. <10% (+) 15% (+)

negative 21%(+) <10% (+) 26.6% (++) 15.5% (+) 10.4 (+) 20.3% (+) 21.6% (+) 10% (+) 18.2% (+) 33% (++) 22.2% (+)

" Ratio of cpm× 100 (%) at a 1:10 dilution for preS-positive serum samples. b Super-responder (R+). Increase in preSIAg/HBsAg ratio at the end of follow-up (PR). n.d., not determined. CAH=chronic active hepatitis. IFN=interferon.

54 of monitoring the success of antiviral therapy in these patients_ For these reasons, the preS2Ag/HBsAg and preSIAg/HBsAg ratios were determined before, during and following antiviral therapy in hepatitis B patients. The human antibody responses to MHBs proteins (gp33/ gp36) and LHBs proteins (p39/gp42) were analyzed and quantified in patients who recovered spontaneously from hepatitis B or responded well to therapy by using the western immunoblotting technique followed by densitometric scanning of blots. Finally, serum HBV-DNA was detected by Polymerase Chain Reaction, an extremely sensitive method for assessing complete virus clearance from the bloodstream. The results of this study show several patterns of response to therapy, first in the antibody response to MHBs and LHBs proteins in complete responders and second in the preS1Ag expression in responders with persistent HBsAg.

L Antibody response to MHBs and LHBs proteins in patients with AH-B Antibodies against the MHBs protein gp33 and against the LHBs protein gp42 were found to be predominant in patients who recovered spontaneously from AH-B_ This suggests that these two monoglycosylated forms of MHBs and LHBs proteins may be more largely expressed on the surface of circulating HBV/HBs particles and/or may be more immunogenic in vivo than the MHBs protein gp36 and the LHBs protein p39_ Such antibodies to MHBs (anti-MHBs) and LHBs (anti-LHBs) proteins appeared only during the late phase of recovery, just before effective production of persistent conformational antibodies to the major small surface (SHBs) antigen (anti-HBs). AntiMHBs and anti-LHBs were never detected by immunoblot early in the course of hepatitis B when patients were HBsAg positive in serum_ This discrepancy with other reports (23-25) is probably due to the nature of antigenic probes used to detect the antibody responses. The antibodies directed against the MHBs proteins (gp33/gp36) and the LHBs proteins (p39/gp42) of the HBV envelope are probably different from antibodies which only recognize a limited sequence of preS2 or preSl regions. Indeed, anti-MHBs and anti-LHBs may correspond to neutralizing antibodies directed against conformational preS epitopes (25). These antibodies may be more important for termination of virus replication in acute infection than antibodies to short preS sequences. II. Effica O, of early IFN therapy in patients with protracted hepatitis B As we have previously suggested (9), the present data confirm that early administration of IFN therapy in patients with protracted hepatitis B is very effective m

M_-A. PETIT et al_ preventing chronicity. Seroconversion to both anti-HBe and anti-HBs associated with loss of serum HBV D N A detected by PCR was observed in approximately 70% of treated patients, reflecting complete elimination of the virus. Patients who failed to respond to IFN therapy were those with elevated baseline preSlAg/HBsAg ratios (>20%) in their serum. This suggests that high serum preS1Ag levels, which are usually related to the presence of virions (8), may be an indicator of non-response to antiviral therapy. Unfortunately, the small number of patients made confirmation impossible. The observation that one patient given placebo recovered fully without producing antibodies to MHBs and LHBs proteins, but only antibodies to SHBs antigen, indicates that complete HBV clearance may occur in the absence of anti-MHBs and anti-LHBs antibodies. Nevertheless, the development of antibody response to the three HBs proteins may contribute to rapid and efficient eradication of the virus during acute HBV infection. As pointed out, most patients with protracted hepatitis B who responded well to IFN therapy with the loss of HBsAg, preS2Ag and preS1Ag developed a strong antibody response to MHBs and LHBs proteins prior to the appearance of anti-HBs. Overall, the immune responses to MHBs and LHBs proteins were found to be associated with: (i) a transient but marked exacerbation of hepatitis (rise in ALT), and (ii) a strong proliferative T cell response to the HBV nucleocapsid antigens (HBc and HBe) in relation to the clearance of circulating HBV envelope antigens during acute HBV infection. Conversely, such a T-cell response to HBcAg and HBeAg appeared to be significantly lower during chronic HBV infection (12). It therefore seems possible that a strong HBe/c-specific T cell response may contribute to development of antibody response to MHBs and LHBs proteins and to immune lysis of infected cells with complete elimination of replicating virus (26). These results support the view that defects in the T cell response to nucleocapsid antigens and in antibody responses to MHBs and LHBs proteins could be responsible for persistence of HBV infection.

III. hwreased preS1Ag/HBsAg ratio may predict relapse in chronic hepatitis B patients who have responded to therapy According to the established virologic (loss of HBeAg and HBV DNA) and biochemical (normalization of serum ALT levels) criteria of responses to antiviral treatment, results of this study show that approximately 50% of HBeAg-positive or anti-HBe-positive chronic hepatitis-B patients responded to IFN therapy. Nevertheless, the retrospective determination of serum pre-

PreSIAg/Ab AND RESPONSE TO INTERFERON

S l A g throughout follow-up of these patients indicated that different patterns of response are indeed observed. Persistence or even an increase in the preS1Ag/HBsAg ratio despite normalization of A L T levels and sustained disappearance of HBV D N A from serum (detected by hybridization) following I F N therapy may reflect lowgrade HBV replication, as documented by the results of PCR. In recent studies (27,28), we demonstrated rearrangement in the preS/S and preC/C genes of defective HBV in two chronic HBV carriers after cloning and sequencing. The preSl domain recognized by our M A b F35.25 and involved in the recognition of HBV by hepatocyte receptors has been found to be conserved on the surface of these m u t a n t HBV molecules. Therefore, it would appear that high serum levels of expression of preSl(F35.25)-epitopes (preSIAg/HBsAg ratios >15%) after the serum clearance of HBeAg and HBV D N A tested by dot-blot hybridization in chronic hepatitis B patients undergoing antiviral therapy may be indicative of virus persistence through selection of defective HBV mutants. In accordance with this hypothesis, sucrose gradient analysis and electron-microscopic examination of circulating viral particles revealed the presence of preSl(F35.25)-epitopes on complete virions during the pretreatment period, and on modified particles (diameter - 3 0 to 35 nm) banding at 30% sucrose during the follow-up after treatment (1-2 years later). We previously showed the presence of HBV m u t a n t forms in circulating modified viral particles with an average diameter of 30 n m in a chronic HBV carrier following therapeutic trials (27). Moreover, HBeAg minus HBV m u t a n t s (29-32) were detected in baseline sera of the majority of patients with anti-HBe-positive chronic hepatitis B, reflecting the preexistence of heterogeneity of HBV D N A molecules before treatment. In summary, this study demonstrates the rationale of introducing preS1Ag detection to monitor antiviral therapy. At present, only the PCR tests gave similar results (33), but the routine use of PCR is a problem, especially for quantification_ Moreover, high serum levels of preS l A g indicate the emergence of defective HBV mutants in anti-HBe-positive HBsAg carriers with persistent viral infection (31).

Acknowledgements This work was supported by grants from the Institut National de la Sant6 et de la Recherche M6dicale (I.N.S.E_R.M.) and the Association Claude Bernard, France.

References 1. Tiollais P, Pourcel C, Dejean A. The hepatitis B virus. Nature 1985; 317: 489-5.

55 2. Machida A, Kishimoto S, Ohnuma H, et al. A polypeptide containing 55 amino acid residues coded by the preS region of hepatitis B virus deoxyribonucleic acid bears the receptors for polymerized human as well as chimpanzee albumins. Gastroenterology 1984; 86: 910-8. 3_ Neurath AR, Kent SBH, Strick N, Parker K. Identification and chemical synthesis of a host cell receptor binding site on hepatitis B virus. Cell 1986; 46: 429-36. 4. Pontisso P, Petit MA, Bankowski MJ, Peeples ME. Human liver plasma membranes contain receptors for the hepatitis B virus PreS1 region and, via polymerized human serum albumin, for the PreS2 region. J Virol 1989; 63: 1981-8. 5. Petit MA, Dubanchet S, Capel F+ V6et P, Dauguet C, Hauser P. HepG2 cell binding activities of different hepatitis B virus isolates: inhibitory effect of anti-HBs and anti-preSl(21-47). Virology 1991; 180: 483-91. 6. Petit MA, Dubanchet S, Capel F. A monoclonal antibody specific for the hepatocyte receptor binding site on hepatitis B virus. Mol lmmunol 1989; 26: 531-7. 7. Petit MA, Strick N, Dubanchet S, Capel F, Neurath AR. Inhibitory activity of monoclonal antibody F35_25 on the interaction between hepatocytes (HepG2 cells) and preSl-specific ligands. Mol Immunol 1991; 28: 517-21. 8+ Petit MA, Zoulim F, Capel F, Dubanchet S, Dauguet C, Tr~po C. Variable expression of preSl antigen in serum during chronic hepatitis B virus (HBV) infection: an accurate marker for the level of HBV replication. Hepatology 1990; I l: 809-14. 9, Tr6po C, Chemin I, Petit MA, et al. Possible prevention of chronic hepatitis B by early interferon therapy. J Hepatol 1990; ll: $95 $99. 10, Causse X, Zoulim F, Godinot H, et al. Chronic hepatitis B therapy with simultaneous steroids and alpha interferon followed by interferon alone [Abstract]. Gastroenterology 1990; 98: A575. 11. Tr6po C, Rougier P, Bizollon T, et al. Combination therapy with ARA-AMP and interferon of chronic active hepatitis BInterim analysis of an ongoing study [Abstract]. J. Hepato[ 1991; 13 (suppl. 1): $3. 12. Ferrari C, Penna A, Bertoletti A, Valli A, et al. Cellular immune response to hepatitis B virus-encoded antigens in acute and chronic hepatitis B virus infection J lmmunol 1990; 145: 3442-9. 13. Bianchi L, De Groote J, Desmet VJ. Acute and chronic hepatitis revisited: review by an international group. Lancet 1977; ii: 914-9. 14. Hantz O, Ooka T, Vitvitski L, Pichoud C, Tr6po C. Comparison of properties of woodchuck hepatitis virus and human hepatitis B virus endogenous DNA polymerase. Antimicrob Agents Chemother 1984; 25: 242-6. 15. Cova L, Lambert V, Chevalier A, et al. Evidence for the presence of duck hepatitis B virus in wild migrating ducks. J Gen Virol 1986; 67: 537-47. 16. Tr6po C, Magnius LO, Schaefer RA, Prince AM. Detection of e antigen and antibody: correlations with hepatitis B surface and hepatitis B core antigens, liver disease, and outcome in hepatitis B infections. Gastroenterology 1976; 71: 804-8_ 17. Chemin I, Baginski I, Petit MA, et al. Correlation between HBV DNA detection polymerase chain reaction and preSl antigenemia in symptomatic and asymptomatic hepatitis B virus infection. J Med Virol 1991; 33: 51-8. 18. Li J, Tong S, Vitvitski L, Zoulim F, Tr6po C. Rapid detection and further characterization of infection with hepatitis B virus variants containing a stop codon in the distal preC region. J Gen Virol 1990; 71: 1993-8_ 19. Petit MA, Capel F, Riottot MM, Dauguet C, Pillot J. Antigenic mapping of the surface proteins of infectious hepatitis B virus particles. J Gen Virol 1987; 68: 2759-67. 20. Petit MA, Capel F, Riottot MM, Pillot J. Epitope analysis of hepatitis B virus envelope glycoprotein complex by monoclonal antibodies. In: Zuckerman AJ, ed. Viral hepatitis and liver disease. New York: Alan R. Liss Inc., 1988: 617-21.

56

M.-A. PETIT et al.

21. Petit MA, Maillard P, Capel F, Pillot J. Immunochemical structure of the hepatitis B surface antigen vaccine. II_ Analysis of antibody responses in human sera against the envelope proteins. Mol Immunol 1986; 23:511-23. 22. Michel ML, Pontisso P, Sobczak E, Malpiece Y, Streeck RE, Tiollais P. Synthesis in animal cells of hepatitis B surface antigen particles carrying a receptor for polymerized human serum. Proc Natl Acad Sci USA 1984; 81: 7708-12. 23. Budkowska A, Riottot MM, Dubreuil P, Lazzizi Y, Br6chot C, Sobczak E, Petit MA, Pillot J. Monoclonal antibody recognizing preS2 epitope of hepatitis B: Characterization of preS2 epitope and anti-preS2 antibody. J Med Virol 1986; 20: 111-25.

24_ Budkowska A, Dubreuil P, Maillard P, Poynard T, Pillot J. A biphasic pattern of anti-preS responses in acute hepatitis B virus infection. Hepatology 1990; 12: 1271-7. 25. Alberti A, Cavalletto D, Chemello L, et al. A. Fine specificity of human antibody response to the preSI domain of hepatitis B virus. Hepatology 1990; 12: 199-203. 26. Milich DR. Genetic and molecular basis for T- and B-cell recognition of hepatitis B viral antigens. Immunol Rev 1987; 99: 71-103. 27. Tran A, Kremsdorf D, Capel F, et al. Emergence of and take-

28.

29.

30.

31.

32.

33.

over by hepatitis B virus (HBV) with rearrangements in the preS/S and preC/C genes during chronic HBV infection. J Virol 1991; 65: 3566-74. Gerken G, Kremsdorf D, Capel F, et al. Hepatitis B defective virus with rearrangements in the preS gene during chronic HBV infection. Virology 1991; 183: 555-65. Carman WF, Jacyna MR, Hadziyannis S, et al. Mutation preventing formation of hepatitis B e antigen in patients with chronic hepatitis B infection. Lancet 1989; ii: 588-91_ Brunetto MR, Stemler M, Bonino F, et al. A new hepatitis B virus strain in patients with severe anti-HBe positive chronic hepatitis B..I Hepatol 1990; 10: 258-61. Okamoto H. Yotsumoto S, Akahane Y, et al. Hepatitis B viruses with precore region defects prevail in persistently infected hosts along with seroconversion to the antibody against e antigen. J Virol 1990; 64: 1298-1303. Tong S, Li J, Vitvitski L, Tr6po C. Active HBV replication in the presence of anti-HBe is associated with viral variants containing an inactive preC region. Virology 1990; 176: 596-603. Carman WF, Dourakis S, Karayannis, et al. Incidence of hepatitis B viraemia, detected using the polymerase chain reaction, after successful therapy of hepatitis B virus carriers with interferon-o:. J Med Virol 1991; 34: 114-8.