CARBOHYDRATE-MEDIATED COUPLING OF PEPTIDES TO IMMUNOGLOBULINS

COUPLAGE DE PEPTIDES A DES IMMUNOGLOBULINES PAR L'INTERMEDIAIRE DE GLUCIDES

English Abstract

The present invention relates to methods for enzymatically coupling peptides
to immunoglobulin molecules via carbohydrate residues of immunoglobulin
molecules, and to immunoglobulin-carbohydrate-linked peptide ("ICLP")
conjugates produced by such methods. ICLP conjugates have been found to be
superior in eliciting an immune response when compared to unconjugated peptide.

French Abstract

Procédés de couplage enzymatique de peptides à des molécules d'immunoglobuline par l'intermédiaire de résidus glucide de molécules d'immunoglobuline. L'invention concerne également des conjugués peptide-immunoglobuline liés par glucide ("ICLP"), obtenus par ces procédés. Les conjugués ICLP apparaissent plus aptes à susciter une réponse immunitaire que les peptides non conjugués.

Claims

-48-

WE CLAIM:

1. A method of conjugating a peptide to an
immunoglobulin molecule via a carbohydrate residue of
the immunoglobulin molecule which comprises the steps
of:
(a) enzymatically oxidizing the carbohydrate
residue of the immunoglobulin molecule;
(b) reacting the oxidized carbohydrate residue
with an amino group of the peptide; and
(c) stabilizing the reaction product of step (b)
by reaction with a reducing agent.

2. The method of claim 1, wherein the
carbo-hydrate residue is galactose or galactosamine, the
galactose or galactosamine being oxidized by galactose
oxidase.

3. The method of claim 2, wherein the
carbo-hydrate residue is galactose.

4. The method of claim 3, wherein the galactose
is covalently linked to a sialic acid residue.

5. The method of claim 4, further comprising a
step of enzymatically removing the sialic acid residue
linked to the galactose residue prior to the enzymatic
oxidation of the galactose residue.

6. The method of claim 5, wherein the sialic
acid is enzymatically removed by neuraminidase.

7. The method of claim 6, wherein the peptide
comprises an immunogenic peptide.

-49-


8. The method of claim 7, wherein the
immuno-genic peptide comprises a B cell epitope.

9. The method of claim 7, wherein the
immuno-genic peptide comprises a helper T cell epitope.

10. The method of claim 6, wherein the peptide
comprises the IL-1 epitope.

11. The method of claim 6, wherein the peptide
comprises the tetanus toxoid epitope.

12. The method of claim 9, wherein the peptide is
covalently linked at an amino group to the amino acid
sequence Ala-Ala-Ala-Leu before it is conjugated to the
carbohydrate residue, such that the Ala-Ala-Ala-Leu
sequence occurs between the peptide and the
carbo-hydrate residue upon conjugation.

13. The method of claim 6, wherein the reducing
agent comprises pyridine borane.

14. The method of claim 1, wherein the peptide
comprises an immunogenic peptide.

15. The method of claim 14, wherein the
immuno-genic peptide is a B cell epitope.

16. The method of claim 14, wherein the
immuno-genic peptide is a helper T cell epitope.

17. A purified immunoglobulin-carbohydrate-linked-peptide
conjugate derivable from the method of claim 1.

-50-


18. A purified immunoglobulin-carbohydrate-linked-peptide
conjugate derivable from the method of claim 6.

19. The purified immunoglobulin-carbohydrate-linked-peptide
conjugate of claim 18, wherein the peptide is a B cell epitope.

20. The purified immunoglobulin-carbohydrate-linked peptide
conjugate of claim 18, wherein the peptide is a helper T cell epitope.

21. The purified immunoglobulin-carbohydrate-linked peptide
conjugate of claim 18, wherein the peptide is the IL-1 epitope.

22. The purified immunoglobulin-carbohydrate-linked peptide
conjugate of claim 18, wherein the peptide is the tetanus toxoid epitope.

23. The purified immunoglobulin-carbohydrate-linked peptide
conjugate of claim 17, wherein the peptide is a B cell epitope or a helper T cell epitope.

24. A purified immunoglobulin molecule comprising
(a) an immunoglobulin molecule and (b) a peptide,
wherein the immunoglobulin is linked to the peptide via
a carbohydrate residue of the immunoglobulin molecule
and wherein the peptide comprises a B cell epitope or a
helper T cell epitope.

25. The purified immunoglobulin molecule of claim
24, wherein the peptide is a B cell epitope.

26. The purified immunoglobulin molecule of claim
24, wherein the peptide is a helper T cell epitope.


-51-

27. A vaccine comprising an effective amount of
the purified immunoglobulin-carbohydrate-linked peptide
conjugate of claim 19 and a suitable carrier.

28. A vaccine comprising an effective amount of
the purified immunoglobulin-carbohydrate-linked peptide
conjugate of claim 20 and a suitable carrier.

29. A vaccine comprising an effective amount of
the purified immunoglobulin-carbohydrate-linked peptide
conjugate of claim 23 and a suitable carrier.

30. A vaccine comprising an effective amount of
the purified immunoglobulin molecule of claim 24 and a
suitable carrier.

31. A method of enhancing an immune response to a
pathogen comprising administering an effective amount
of the purified immunoglobulin-carbohydrate-linked-peptide
conjugate of claim 19 and a suitable carrier.

32. A method of enhancing an immune response to a
pathogen comprising administering an effective amount
of the purified immunoglobulin-carbohydrate-linked-peptide
conjugate of claim 20 and a suitable carrier.

33. A method of enhancing an immune response to a
pathogen comprising administering an effective amount
of the purified immunoglobulin-carbohydrate-linked-peptide
conjugate of claim 23 and a suitable carrier.

34. The method of claim 7, wherein the
immuno-genic peptide comprises a cytotoxic T cell epitope.

-52-

35. The method of claim 14, wherein the
immuno-genic peptide is a cytotoxic T cell epitope.

36. The purified immunoglobulin-carbohydrate-linked
peptide conjugate of claim 18, wherein the
peptide is a cytotoxic T cell epitope.

37. The purified immunoglobulin-carbohydrate-linked
peptide conjugate of claim 17, wherein the peptide
is a cytotoxic T cell epitope.

38. A purified immunoglobulin molecule comprising
(a) an immunoglobulin molecule and (b) a peptide,
wherein the immunoglobulin is linked to the peptide via
a carbohydrate residue of the immunoglobulin molecule
and wherein the peptide comprises a cytotoxic T cell
epitope.

Description

CA 02227326 l998-0l-l~

W096/36357 PCT~S96/06756
--1--


DescriPtion
CARBOHYDRATE-MEDIATED COUPLING OF PEPTIDES TO
IMMUNOGLOBULINS
1. INTRODUCTION
The present invention relates to methods for
coupling peptides to immunoglobulin molecules via
galactose or other sugar residues, and to immuno-
globulin-carbohydrate-linked peptide ("ICLP") con-
jugates produced by such methods. ICLP conjugates havebeen found to be superior in eliciting an immune
response when compared to unconjugated peptide.
2. BACKGROUND OF THE lN V~'N'l'lON
2.1. COUPLING PEPTIDE TO CARRIER
TO ENHANCE IMMUNOGENICITY
Isolated peptides are frequently too small and/or
too unstable to elicit an immune response by them-
selves, and for this reason, a number of methods have
been developed for linking peptides to larger "carrier"
molecules. Enhancement of in vivo immunogenicity of
such peptides depends upon the structure of the carrier
to which they are coupled, as well as the type of
intermolecular cross-linking between the peptide and
its carrier.
One method for linking a peptide to a carrier is
chemical conjugation. Chemical conjugates of pro-
teinaceous carriers with synthetic peptides have been
observed to elicit humoral and cellular anti-peptide
immune responses in laboratory ~n; ~
However, although more than 300 chemical cross-
linkers with various reactivities have been developed
over the past years, only a few are currently used to
generate biologically active conjugates (Wong, 1991, in
"Chemistry of Protein Conjugation And Cross-Linking",
CRC Press, Inc., Boca Raton, Florida; Means and
Freeney, 1971, Bioconj. Chem. 1:2). A major drawback

CA 02227326 l998-Ol-l~
W096/36357 PCT~S96/06756



of chemical cross-linkers is the generation of new
epitopes in the context of the carrier/cross-
linker/peptide complex. Furthermore, because chemical
cross-linkers typically lack specificity toward parti-
cular groups, additional steps may be required to pro-
tect cognate structures other than sites targeted for
linkage (Freeney, 1987, Int. J. Peptide Prot. Res.
29:145).
The advent of molecular biology has allowed for
the preparation of chimeric molecules in which mini-
genes encoding biologically active peptides are expres-
sed in carrier genes, to result in recombinant expres-
sion of peptide linked to carrier in a single molecule
(Leclerc, 1994, Intern. Rev. Immunol. 11:103).
Because strong immune responses may be elicited
against determinants borne by the carrier portion,
rather than the incorporated peptide (C~ ~ota et al.,
1992, Nature 356:799), genes encoding self proteins
have been genetically engineered to carry foreign
epitopes (Lanza et al., 1993, Proc. Natl. Acad. Sci.
U.S.A. 90:11683).
For example, expression of microbial B and T cell
epitopes in the CDR3 loop of self immunoglobulins has
been used to create chimeric molecules which were found
to elicit epitope-specific cellular and humoral immune
responses in vivo (Zaghouani et al., 1993, Science
259:224). Genetically antigenized immunoglobulin
("Ig") carrying a CD4 epitope (HA 110-120) from hemag-
glutinin ("HA") of influenza PR8 A virus, namely Ig-HA,
was able to elicit specific and efficient immune
responses in mice (Id . ) . This was correlated with high
amounts of viral peptides released from the chimeric
immunoglobulins into antigen presenting cells ("APCs")
and presented efficiently to specific T helper cells
(Brumeanu, 1993, J. Exp. Med.178:1795).

CA 02227326 1998-01-1~

WO 96/3635? PCT/US96/067S6
--3--


2.2. CONJUGATION VIA GLYCOSIDIC LINKAGES
Radioactive labelling of glycoproteins has been
accomplished by conjugating radioactive cysteine methyl
~ ester to aldehydes produced on penultimate galactose
residues generated by treating glycoproteins with neura-
minidase (to remove terminal sialic acid residues) and
then with galactose oxidase (to generate aldehyde func-
tional groups) (Mitchell et al., 1984, Arch. Biochem.
Biophys. 229:544-554). A similar method had previously
been used to label proteins with tritium, wherein the
aldehyde groups of the glycoproteins were reduced with
tritiated borohydride (Morell and Ashwell, 1972,
Methods Enzymol. 28:205).
In another example of conjugation via carbohydrate
residues, peroxidized sugars have been used to cross-
link glycoproteins, including some enzymes. The
utility of this chemistry is restricted because cross-
linking disturbs the biological functioning of most
enzymes (Eyzaguirre, 1987, in "Chemical Modification of
Enzymes: Active Site Studies", Ellis Horwood,
Chichester, England).
3. SUMMARY OF THE INVENTION
The present invention relates to methods for
coupling peptides to carbohydrate moieties normally
occurring in immunoglobulin molecules, and to immuno-
globulin-carbohydrate-linked peptide ("ICLP") con-
jugates produced by such methods.
In particular non-limiting embodiments of the
invention, peptides are conjugated to carbohydrate
3 0 residues of an immunoglobulin molecule by enzymatically
oxidizing the carbohydrate residues to produce aldehyde
groups, by reacting the oxidized carbohydrate residues
with the peptides such that covalent attachments
between the residues and peptides are formed, and by
35 stabilizing the bond between the residues and peptides

CA 02227326 1998-01-1~

W096/36357 PCT~S96/06756
-4-

by reduction using appropriate reducing agents. In a
preferred embodiment of the invention, the enzymati-
cally oxidized carbohydrate residues are galactose
residues and galactose oxidase may be used for enzyma-
tic oxidation. It is preferred that the enzymaticoxidation of galactose residues is preceded by enzyma-
tic removal of terminal sialic acid residues attached
to galactose residues. Neuraminidase may be used for
the enzymatic desialylation.
The present invention is based, at least in part,
on the discovery that when a peptide corresponding to
amino acid residues 110-lZ0 of the hemagglutinin of
influenza PR8 A virus (SEQ ID NO:7) was conjugated to
immunoglobulin molecules via carbohydrate residues, the
resulting ICLP conjugates were observed to activate HA
110-120 specific T cell hybridoma cells as efficiently
as influenza PR8 virus, at levels 40-100 fold higher
than the synthetic HA 110-120 peptide itself.
Utilizing immunoglobulin molecules as carriers for
synthetic peptides offers a number of advantages. Use
of ICLP conjugates may not only prolong the half-life
of the peptide, but may also, via binding of Fc regions
of the immunoglobulin molecule to cell surface recep-
tors, recruit elements of the ; ~ system so as to
augment and improve the efficiency of the overall
immune response.
As another advantage, the specificity of enzymes
used to link peptide with immunoglobulin substantially
reduces the generation of by-products and the creation
of undesirable neodeterminants.

4. DESCRIPTION OF THE FIGURES
FIGURE 1: Protocol for the synthesis of immuno-
globulin-carbohydrate-linked-peptide ("ICLP") con-
jugates, which describes a preferred embodiment of the
conjugation method of the present invention. Sugar

CA 02227326 1998-01-1~

WO 96136357 PCT/US96/06756
--5--


moieties of mouse and human immunoglobulins are
~ depleted of N-acetylneuraminic acid (NANA, sialic acid)
using neuraminidase from Arthrobacter ureafaciens
(Powell and Varki, 1993, in "Sialidases", Green
Publishing and John Wiley & Sons, New York, 17.12.1-
17.12.8) and Clostridium ~erfrinqens, and the adjacent
galactose residues are subsequently oxidized at the
carbon-6 position using galactose oxidase (Cleveland et
al., 1975, Biochem. 14:1108). The reductive alkyl-
ation is favored between the aldehyde group generatedenzymatically on the galactose and the ~ amino group of
the synthetic peptide. The synthetic peptide contains
a N-terminal site for cathepsin E (Ala-Ala-Ala-Leu; SEQ
ID N0:15) which has been artificially introduced to
facilitate quick release of the peptides into the lyso-
somal compartment of the antigen processing cells.
Schiff bases formed between galactose and peptides are
stabilized by reduction with pyridine borane
(Cabacungan et al., 1982, Anal. Biochem.124:272).
FIGURE 2: Analysis of mouse IgG-carbohydrate-
linked-HA conjugates by SDS -PAGE and Western blot.
Mouse monoclonal IgG1 was (conjugated, via carbohydrate
residues, to HAC110-120 synthetic peptide, dialyzed
against PBS in bags of 100,000 MWC0 and aliquots were
analyzed by SDS-PAGE under nonreducing and reducing
conditions as described in Section 6. Lanes 1 and 5,
respectively, represent Coomassie staining of the non-
reduced and reduced IgGl(control), and lanes 2 and 6,
respectively, show Coomassie staining of nonreduced and
reduced IgG-carbohydrate-linked-HA conjugates. Western
blots of the nonreduced IgG and IgG-carbohydrate-
linked-HA that were developed with rabbit anti-HA 110-
120 antibodies are shown in lanes 3 and 4,
respectively, and the reactivity of these with the
reduced IgG and IgG-carbohydrate-linked-HA are shown in
lanes 7 and 8, respectively. Identification of the

CA 02227326 l998-Ol-l~

W096/36357 PCT~S96/06756
-6-


light and heavy chains of the reduced IgG and IgG-
carbohydrate-linked-HA conjugate was performed using
Western blots developed with rabbit anti-murine yl and
K chain antibodies; lanes 9 and 11 show, respectively,
the heavy and light chains o~ the reduced IgG, and
lanes 10 and 12 show, respectively, the heavy and light
chains of the IgG-carbohydrate-linked-HA conjugate. As
can be seen in lane 8, the heavy chains of mouse IgG
were coupled to HAC110-120 peptide and their molecular
weight was found to be slightly increased (lanes 6 and
8).
FIGURE 3: Chromatographic removal of residual,
unconjugated HAC110-120 peptide from the ICLP con-
jugates. A Superose-6 gel filtration column was
previously calibrated with molecular weight markers
(Pharmacia), and then mouse IgG-carbohydrate-linked-HA
and IgM-carbohydrate--linked-HA preparations were
chromatographed as described in Section 6. Major peaks
represent either native mouse IgG and IgM or IgG-carbo-
hydrate-linked-HA and IgM-carbohydrate-linked-HA con-
jugates. The late peak that eluted at 80 minutes, as
indicated on the chromatograms of ICLP conjugàtes,
represents residual free peptide. The peptide
identification in the peak eluted at 80 minutes was
confirmed by IRIA (competitive inhibition radio-
immunoassay) and RP-HPLC (reverse-phase HPLC).
FIGURE 4: Specificity of attachment of HAC110-l20
peptide to the sugar moiety of the immunoglobulins.
Western blot analysis of the reduced IgG-carbohydrate-
linked-HA conjugates, before and after treatment with
PGNase F, lanes 1 and 2, respectively, was developed
with rabbit anti-HA 110-120 antibodies as described in
Section 6. The lack of reactivity of rabbit anti-HA
110-120 antibodies to the heavy chains of the PGN-ase F
treated conjugate indicates removal of the N-linked
oligosaccharide/HAcllo-l2o complex from the heavy

CA 02227326 l998-Ol-l~

W096/36357 PCT~S96/06756
--7--


chains of mouse IgG-carbohydrate-linked-HA conjugate
(lane 2).
FIGURE 5: Estimation of the degree of coupling of
HAc110-120 peptide to mouse IgG. Mouse IgG-carbo-
hydrate-linked-HA conjugate was prepared, purified and
hydrolyzed with PGNase as described. The amount of the
enzymatically detached HAC110-120 peptide from the con-
jugate (open squares) was estimated by IRIA as a
measure of percent inhibition of binding of rabbit
anti-HA110-120 antibodies to plates coated with BSA-HA
conjugate. Dotted lines show 50 percent inhibition
obtained with either HAC110-120 peptide detached from
the conjugate or HA110-120 synthetic peptide (cali-
bration). The amount of HAC110-120 peptide (0.9ng)
detected in 75 ~l, as used in IRIA (competitive inhi-
bition radioimmunoassay), was related to the cor-
responding amount of mouse IgG1 (86.4ng) found in 75
~l. Since 16.9:1 represents a molar ratio of 1:1
between heavy chain of IgG and HAC1l0-120 peptide,
and most of the peptides were found attached to the
oligosaccharide chains of the heavy chain, the number
of peptide units per heavy chain was calculated to be
5.68. This corresponds to an average of 11.4 peptides
per molecule of IgG. Data points on the graph
represent the mean of triplicate samples + SD.
FIGURE 6: Activation of HA110-120 specific, LD1-
24 T hybridoma cells by mouse IgG-carbohydrate-linked-
HA conjugates. 2PK3 (APC) cells were incubated with
graded amounts of HA110-120 synthetic peptide, NP147-
161 synthetic peptide (control), W-inactivated PR8
virus, genetically antigenized Ig-HA, genetically
antigenized Ig-NP (control), enzymatically antigenized
mouse IgG-carbohydrate-linked-HA and IgM-carbohydrate-
linked-HA conjugates, and their appropriate controls.
IgG-carbohydrate-linked-HA and IgM-carbohydrate-linked-
HA conjugates were previously rendered free of uncon-


CA 02227326 1998-01-1~

W096/36357 PCT~S96/06756
-8-


jugated HACll0-l20 peptide by extensive dialysis fol-
lowed by size exclusion chromatography. The ability of
each of the antigens to specifically activate the LD1-
24 T hybridoma cells was compared at 50~ activation as
indicated by dotted lines. Data points on the graph
represent the mean of quadruplicate samples + SD.
FIGURE 7: Western blot analysis of enzymatically
antigenized mouse and human immunoglobulin. Various
isotypes of affinity purified mouse monoclonal and
human myeloma IgG were glycosidically coupled with
HACll0-l20 antibodies as described. It should be noted
that residual unconjugated HACll0-l20 peptide was not
chromatographically removed from these particular con-
jugates. Mouse IgG-carbohydrate-linked-HA (lane 1) and
human IgG-carbohydrate-linked-HA (lane 3), mouse IgM-
carbohydrate-linked-HA (lane 5), human IgM-carbo-
hydrate-linked-HA (lane 7) and human IgA-carbohydrate-
linked-HA (lane 9) were analyzed in parallel with their
appropriate controls: mouse IgG (lane 2), human IgG
(lane 4), mouse IgM (lane 6), human IgM (lane 8), and
human IgA (lane 10).
FIGURE 8. (a) T cell activation as a function of
the antigen carrier molecule for HA 110-120. Open
circles represent HA110-120 synthetic peptide; clear
diamonds represent HA110-120 comprised in the CDR3 loop
of an immunoglobulin; solid diamonds represent an IGP
conjugate of the HA 110-120 peptide; and open triangles
represent the HA110-120 peptide comprised in bromelain
released HA protein. (b) HA110-120 immunopotency in the
context of various antigen carriers (symbols as set
forth in (a)).
FIGURE 9. T cell activation indices for HA110-120
comprised in various carriers and exposed to (a)
irradiated (empty bars) or fixed (solid bars) APCs; (b)
chloroquine-treated APCs; or (c) fixed APCs together

CA 02227326 1998-01-1~

W096/36357 PCT~S96/06756
_g_


with anti-Fc gamma receptor and anti-class I
~ inhibitors.
FIGURE 10. Diagram of model for antigen
- presentation by APC to a T helper cell.

5. DETAILED DESCRIPTION OF THE INVENTION
For purposes of clarity, and not by way of
limitation, the detailed description of the invention
is divided into the following subsections:
(i) peptides suitable for conjugation;
(ii) immunoglobulins suitable for
conjugation;
(iii) enzymatic coupling of peptide to
immunoglobulin; and
(iv) utility of the invention.

5.1. PEPTIDES SUITABLE FOR CONJUGATION
Virtually any peptide may be conjugated to an
immunoglobulin according to the invention. Peptides
for use according to the invention comprise at least 2
and preferably at least five, amino acid residues and
may comprise immunogenic epitopes of antigens; suitable
antigens include, but are not limited to, antigens
associated with pathogens, tumor cells, or "non-self"
antigens with respect to a particular individual.
Peptides may be biologically active themselves (for
example, growth factors, toxins, immune mediators,
differentiation factors etc.) or be portions of
biologically active proteins.
In non-limiting embodiments, peptides which may be
conjugated to immunoglobulins, according to the inven-
tion, include B cell epitopes. The term "B cell epi-
tope", as used herein, refers to a peptide, including a
peptide comprised in a larger protein, which is able to
bind to an immunoglobulin receptor of a B cell and

CA 02227326 l998-0l-l~
W096/36357 PCT~S96/067S6
--10--


participate in the induction of antibody production by
the B cell.
For example, and not by way of limitation, the
hypervariable region 3 loop ("V3 loop") of the envelope
protein of human immunodeficiency virus ("HIV") type 1
is known to be a B cell epitope. Although the sequence
of this epitope varies, the following consensus
se~uence, corresponding to residues 301-319 of HIV-l
gpl20 protein, has been obtained: Arg-Lys-Ser-Ile-His-
Ile-Gly-Pro-Gly-Arg-Ala-Phe-Tyr-Thr-Thr-Gly-Glu-Ile-Ile
(SEQ ID NO:l).
Other examples of known B cell epitopes which may
be used according to the invention include, but are not
limited to, epitopes associated with influenza virus
strains, such as Trp-Leu-Thr-Lys-Lys-Gly-Asp-Ser-Tyr-
Pro (SEQ ID NO:2), which has been shown to be an
immuno~t ;n~nt B cell epitope in site B of influenza
HA1 hemagglutinin, the epitope Trp-Leu-Thr-Lys-Ser-Gly-
Ser-Thr-Tyr-Pro (H3; SEQ ID NO:3), and the epitope Trp-
Leu-Thr-Lys-Glu-Gly-Ser-Asp-Tyr-Pro (H2; SEQ ID NO:4)
(Li et al., 1992, J. Virol. 66:399-404); an epitope of
F protein of measles virus (residues 404-414; Ile-Asn-
Gln-Asp-Pro-Asp-Lys-Ile-Leu-Thr-Tyr; SEQ ID NO:5;
Parlidos et al., 1992, Eur. J. Immunol. 22:2675-2680);
an epitope of hepatitis virus pre-Sl region, from
residues 132-145 (Leclerc, 1991, J. Immunol. 147:3545-
3552); and an epitope of foot and mouth disease VPl
protein, residues 141-160, Met-Asn-Ser-Ala-Pro-Asn-Leu-
Arg-Gly-Asp-Leu-Gln-Lys-Val-Ala-Arg-Thr-Leu-Pro (SEQ ID
NO:6; Clarke et al., 1987, Nature 330:381-384).
Still further B cell epitopes which may be used
are known or may be identified by methods known in the
art, as set forth in Caton et al., 1982, Cell 31:417-
427.
In additional embodiments of the invention,
peptides which may be conjugated to immunoglobulins may

CA 02227326 l998-0l-l~

W096l36357 PCT~S96/06756
--11--


be T cell epitopes. The term "T cell epitope", as used
herein, refers to a peptide, including a peptide com-
prised in a larger protein, which is associated with
MHC self antigens and recognized by a T cell and which
functionally activates the T cell.
For example, a Th epitope is recognized by a
helper T cell and promotes the facilitation of B cell
antibody production via the Th cell. Such epitopes are
believed to arise when antigen presenting cells
("APCs") have taken up and degraded a foreign protein.
The peptide products of degradation, which include the
Th epitope, appear on the surface of the APC in associ-
ation with MHC class II antigens.
Further, a CTL epitope is recognized by a cyto-
toxic T cell; when the CTL recognizes the CTL epitopeon the surface of a cell, the CTL may induce lysis of
the cell. CTL epitopes are believed to arise when an
APC has synthesized a protein which is then processed
in the cell's cytoplasmic compartment, leading to the
generation of peptides (including the CTL epitope)
which associate with MHC class I antigens on the sur-
face of the APC and are recognized by CD8+ cells,
including CTL.
For example, and not by way o~ limitation, in~lu-
enza A hemagglutinin (HA) protein bears, at amino acidresidues 110-120, a Th epitope having the amino acid
sequence Ser-Phe-Glu-Arg-Phe-Glu-Ile-Phe-Pro-Lys-Glu
(SEQ ID NO:7).
Other examples of known T cell epitopes include,
but are not limited to, two promiscuous epitopes of
tetanus toxoid, Asn-Ser-Val-Asp-Asp-Ala-Leu-Ile-Asn-
Ser-Thr-Lys-Ile-Tyr-Ser-Tyr-Phe-Pro-Ser-Val (SEQ ID
NO:8) and Pro-Glu-Ile-Asn-Gly-Lys-Ala-Ile-His-Leu-Val-
Asn-Asn-Glu-Ser-Ser-Glu (SEQ ID NO:9; Ho et al., 1990,
Eur. J. Immunol. 20:477-483); an epitope of cytochrome
c, from residues 88-103, Ala-Asn-Glu-Arg-Ala-Asp-Leu-

CA 02227326 1998-01-1~
W096/36357 PCT~S96/06756
-12-

Ile-Ala-Tyr-Leu-Gln-Ala-Thr-Lys (SEQ ID NO:10); an
epitope of Mycobacteria heatshock protein, residues
350-369, Asp-Gln-Val-His-Phe-Gln-Pro-Leu-Pro-Pro-Ala-
Val-Val-Lys-Leu-Ser-Asp-Ala-Leu-Ile (SEQ ID NO:ll;
Vordermir et al., Eur. ~. Immunol. 24:2061-2067); an
epitope of hen egg white lysozyme, residues 48-61, Asp-
Gly-Ser-Thr-Asp-Tyr-Gly-Ile-Leu-Gln-Ile-Asn-Ser-Arg
(SEQ ID NO:12; Neilson et al., 1992, Proc. Natl. Acad.
Sci. U.S.A. 89: 7380-7383; an epitope of Streptococcus
A M protein, residues 308-319, Gln-Val-Glu-Lys-Ala-Leu-
Glu-Glu-Ala-Asn-Ser-Lys (SEQ ID NO:13; Rossiter et al.,
1994, Eur. J. Immunol. 24:1244-1247); and an epitope of
Staphylococcus nuclease protein, residues 81-100, Arg-
Thr-Asp-Lys-Tyr-Gly-Arg-Gly-Leu-Ala-Tyr-Ile-Tyr-Ala-
Asp-Gly-Lys-Met-Val-Asn (SEQ ID NO:14; de Magistris,
1992, Cell 68:1-20). Still further Th epitopes which
may be used are known or may be identified by methods
known in the art.
As another example, and not by way of limitation,
PR8 influenza virus nucleoprotein bears, at amino acid
residues 147-161, a CTL epitope having the amino acid
sequence Thr-Tyr-Gln-Arg-Thr-Arg-Ala-Leu-Val-Arg-Thr-
Gly-Met-Asp-Pro (SEQ ID NO:16).
Other examples of known CTL epitopes include, but
are not limited to, those discussed in Rotzschke et
al., 1991, Immunol. Today 12:447-455. Still further CTL
epitopes which may be used are known or may be identi-
fied by methods known in the art.
If the peptide to be conjugated to immunoglobulin
comprises a T cell epitope, it may be desirable to
provide a means for releasing the T cell epitope from
the immunoglobulin so as to facilitate appropriate pro-
cessing by an antigen presenting cell ("APC"). For
example, it may be desirable to incorporate the T cell
epitope into a larger peptide which comprises a site
susceptible to cleavage by an enzyme typically present

CA 02227326 1998-01-1~

W096/36357 PCT~S96/067~6
-13-


in lysosomes of APCs. As a specific example, the lipo-
philic quadruplet Ala-Ala-Ala-Leu (SEQ ID N0:15), that
contains the cleavage site for cathepsins (Yonezawa
et al., 1987, Arch. Biochem. Biophys 256:499), may be
added to the T cell epitope. It is known that
lysosomal cathepsins play an important role in pro-
cessing of exogenous molecules.
It has also been observed that class II antigens
conjugated to immunoglobulin appear to elicit an immune
response when exposed to fixed antigen presenting
cells, consistent with immunogenicity in the absence of
antigen processing (see, for example, Section 7,
below). Accordingly, class I and class II antigen
epitopes may be utilized according to the invention.
In additional embodiments of the invention, pep-
tides conjugated to immunoglobulins may be cytokines.
For example, IL-1 may be conjugated and delivered to
the immune cells to provide stimulatory effect. In
another example, IL-2 may be conjugated to immuno-
globulins having recognition sites for antigens present
on tumor cells to effect delivery of the IL-2 to tumor
cells for purpose of inhibiting their growth.
In additional embodiments, other pharmaceutical
agents directed against various pathogens such as
virus, bacteria, or fungi may be conjugated to immuno-
globulins. By conjugating the pharmaceutical agents to
immunoglobulins having recognition sites for antigens
present on the pathogens to which the agents are direc-
ted against, targeted delivery of those agents may be
effected. Toxins, such as tetanus toxoid, can also be
conjugated to appropriate immunoglobulins for targeted
delivery.
It should be noted that the conjugation method of
the present invention may also be used to couple to an
immunoglobulin molecule not just peptides, but other
compounds. For example, the pharmaceutical agents to

CA 02227326 1998-01-1~
W096/36357 PCT~S96/067~6
-14-


be coupled do not necessarily have to be peptides, as
long as they possess certain reactive sites, whether
they are present naturally or introduced into the
agents for specific purpose of conjugation, for
reacting with the aldehyde groups of the oxidized car-
bohydrate residues to form covalent linkages. One such
reactive site is provided by the presence of an amino
group.
Peptides for use according to the invention may be
purified from natural sources, or may be recombinantly
and/or chemically synthesized. They may contain amino
acid analogs, and may be detectably labelled.

5.2. IMMUNOGLOBULINS SUITABLE FOR CONJUGATION
Immunoglobulins that may be used according to the
invention include human and non-human mammalian immuno-
globulins, as well as immunoglobulins prepared by com-
bining portions of human and non-human ~ lian
immunoglobulin. The immunoglobulins that may be used
in the invention also include those engineered using
recombinant technology. Nonlimiting examples of such
recombinant immunoglobulin may be a human immuno-
globulin from which the CDR regions are removed and
replaced with CDR regions of a murine antibody or a
human immunoglobulin from which the variable regions
are removed and replaced with the variable regions of a
murine antibody. The immunoglobulins may be of any
class, including IgG, IgM, IgD, IgE, and IgA. ICLP
conjugates prepared using immunoglobulins of the IgM
class may exhibit somewhat greater immunogenic activity
than ICLP conjugates prepared using IgG class immuno-
globulin. Immunoglobulin may be monoclonal or poly-
clonal, and may be obtained from cell cultures or from
serum of a human or a non-human mammal.
The term "immunoglobulin", as used herein, gen-
erally refers to molecules comprising both heavy and

CA 02227326 1998-0l-l~

W096/36357 PCT~S96/06756
-15-


light chains. However, if desirable, the conjugating
methods of the invention may be used to link peptides
to either heavy or light chains separately.
Further, peptides may be linked to immunoglobulin
molecules which have particular desirable properties or
activities such as immunoglobulin molecules which are
pinocytosed to the cytoplasm or which are transported
to the nucleus.
As previously set forth in section 5.1, specific
nonlimiting embodiments of the present invention
include conjugation of pharmaceutical agents with
immunoglobulins having recognition sites for antigens
present on cells or pathogens to which the pharma-
ceutical agents are directed. Accordingly, the immuno-
globulins of the present invention include those havingrecognition sites for targeted cells or pathogens.

5.3. ENZYMATIC COUPLING OF PEPTIDE TO IMMUNOGLOBULIN
A schematic diagram depicting a preferred embodi-
ment of the coupling method of the invention is set
forth in Figure 1.
First, immunoglobulin may be treated with a neura-
minidase to remove all or substantially all of the
terminal sialic acid attached to carbohydrate residues
on the immunoglobulin. "Substantially all", as used in
this paragraph, indicates that at least 90 percent of
the sialic acid residues have been removed. The neura-
minidase may be obtained from, for example, Arthro-
bacter ureafaciens, Clostridium perfringens or Vibrio
cholerae, or any source producing a neuraminidase
having alpha 2-6 or alpha 2-3 specificity . In a pre-
ferred, nonlimiting embodiment, neuraminidases from
Arthrobacter ureafaciens and Clostridium perfringens
(50 mU of a 1:1 mixture of the neuraminidases obtained
from the two sources) may be incubated overnight at
37~C with lml of phosphate buffer, pH 6, containing

CA 02227326 1998-01-1~
W096/36357 PCT~S96/06756
-16-

mouse or human immunoglobulins (1 mg) and 5 mM of
CaCl2. Depending on the neuraminidase used, however,
it may not be necessary to include CaC12 in the reac-
tion mixture. The kinetics of the desialylation reac-
tion may be monitored by determining free NANA releasedin solution as described in Warren, 1959, J. Biol.
Chem. 234:1971.
It should be noted that even though neuraminidase
treatment of the immunoglobulin molecule to remove
terminal sialic acids before the enzymatic oxidation of
carbohydrate residues is not necessary for the purposes
of the invention, such treatment is preferred. This is
especially so when the carbohydrate residues to be oxi-
dized are galactose residues. By treating the immuno-
globulin with neuraminidase, terminal sialic acids areremoved, l~n~king those galactose residues linked to
terminal sialic acids and maximizing the number of
sites available for conjugation.
Second, NANA residues released by the desialyla-
tion process may be substantially removed from thedesialylated immunoglobulin composition. Such removal
may be accomplished, for example, by dialysis, affinity
chromatography, or gel filtration. In a specific, non-
limiting embodiment, desialylated immunoglobulins may
be dialyzed against PBS (pH 7) until substantially free
NANA is removed.
Third, carbohydrate residues, such as, in particu-
lar, galactose residues, lln-~ked by desialylation may
be enzymatically oxidized at the C-6 position by galac-
tose oxidase ("GA0"). Further, the carbohydrate resi-
due to be oxidized may be either an internal or a
terminal residue. Preferably, carbohydrate residues to
be oxidized are galactose or galactosamine residues,
which may be oxidized by galactose oxidase. The glu-
cose and glucosamine residues may also be oxidized, forexample, by using glucose oxidase.

CA 02227326 l998-Ol-l~

wos6/36357 PCT~S96/06756
-17-


Fourth, the Schiff bases formed between peptides
and carbohydrate residues may be stabilized by reduc-
tion. For this reaction, reducing agents, including
but not limited to, pyridine borane, sodium boro-
hydride, sodium cyanoborohydride and mercaptoethanolmay be used. In a specific, nonlimiting embodiment,
GA0 (10 U), a reducing agent pyridine borane ("PB"; 40
mM) and peptide (100 fold molar excess) may be added
under continuous stirring for 48 hours at 37~C.
Stabilization of Schiff bases formed between peptides
and carbohydrate residues occurs upon reduction
(Cabacungan et al., 1982, Anal. Biochem. 124:272). The
oxidation and reduction reactions may also be performed
sequentially.
Fifth, after completion of the coupling reaction,
ICLP conjugates may be separated from unreacted reac-
tants by methods such as dialysis, affinity chroma-
tography, HPLC, etc. In a specific, nonlimiting
embodiment, the mixture may be dialyzed against PBS in
bags with 100,000 MWC0 (Spectrapor), and then concen-
trated to 0.1 ml in ultra concentrators of 100,000 MWC0
(S&S).
In particular embodiments, it may be desirable to
conjugate a single species of peptide to an immuno-
globulin molecule. In other embodiments, it may be
desirable to conjugate a diversity of peptides to an
immunoglobulin molecule.
A detailed description of one specific, non-
limiting embodiment of the invention is set forth in
section 6, below.

5.4. UTILITY OF THE I-NV~N'1'10N
The ICLP conjugates of the invention may be used
in a number of commercial, diagnostic, and therapeutic
applications.

CA 02227326 l998-Ol-l~
W096/36357 PCT~S96/06756
-18-


In a first set of embodiments of the invention,
ICLP conjugates may be used in affinity purification
methods of a ligand of the ICLP-comprised paptide. For
example, such a ligand may be a molecule expressed by
recombinant techniques in a bioreactor.
In a second set of embodiments, ICLP conjugates
may be used in detecting or quantitating the presence
or amount of the target antigen of the immunoglobulin
comprised in the ICLP conjugate. For example, the pep-
tides comprised in the ICLP conjugate may be detectablylabelled, and the ICLP conjugate may be exposed to the
target antigen under conditions that permit the binding
of the immunoglobulin comprised in the ICLP conjugate
to its target antigen. Alternatively, the ICLP con-
jugate may be exposed to the target antigen under con-
ditions that permit the binding of the immunoglobulin
comprised in the ICLP conjugate to its target antigen,
and then the ICLP-target antigen complex may be further
reacted to a detectably labelled secondary antibody
directed toward the peptide comprised in the ICLP con-
jugate. Since a number of peptides are comprised in
each ICLP conjugate, the magnitude of the signal pro-
duced by the label would be multiplied.
ICLP conjugates, where the peptide is a B-cell
epitope, may further be used to label B cells. For
example, the peptides comprised in the ICLP may be
detectably labelled, and may be used to quantitate the
number of B cells binding to the particular epitope in
a sample of lymphocytes collected from the subject.
Similarly, such an ICLP conjugate, which need not be
detectably labelled, may be used to test the ability of
a subject to mount a humoral response to the particular
B cell epitope. The inability of lymphocytes of a sub-
ject to produce antibodies after exposure to the ICLP
conjugate may indicate that the subject is not capable
of developing humoral immunity to the epitope. Fur-


CA 02227326 1998-01-1~

WO 96136357 PCTIUS96/06756
--19--


ther, such an ICLP conjugate may be used to collect B
cells which recognize the epitope; for example, the
peptides comprised in the ICLP conjugate may be fluores-
cently labelled, so that the B cells bound to ICLP con-
jugate may be collected by fluorescence-activated cell
sorting.
ICLP conjugates, where the comprised peptide is a
Th cell epitope, may further be used to test the
ability of a subject to mount an immune response to the
particular Th cell epitope. For example, peripheral
blood lymphocytes may be collected from a test subject,
and then, in a st~n~d proliferative assay, may be
exposed to the ICLP conjugate bearing the Th epitope.
The amount of proliferation may then be determined, and
may be compared to the degree of proliferation exhi-
bited by peripheral blood lymphocytes from a control
subject who has not been exposed to the epitope. A
result, in which the amount of proliferation exhibited
by the lymphocytes from the test subject is signifi-
cantly greater than the amount of proliferation exhi-
bited by the lymphocytes from the control subject,
positively correlates with prior exposure of the test
subject to the epitope, and may indicate that the test
subject is or has been infected with a pathogen con-
taining the epitope.
In further embodiments of the invention, ICLPconjugates may be useful in the treatment of a wide
variety of malignancies and viral infections. They are
particularly well suited for treatment of infections by
viruses which upon infection of the host cell cause
expression of viral coat proteins prior to cell death.
In most cases this cellular expression of viral coat
proteins leads to a cell surface form of such proteins.
Examples include but are not limited to the hemag-
glutinin protein complex of influenza virus, the envproteins of murine leukemia virus, the env proteins of

CA 02227326 l998-0l-l~
W096/36357 PCT~S96/06756
-20-


Rous sarcoma virus and the env proteins of HIV. Often
the viral protein expressed by infected cells is the
same viral coat protein which recognizes and binds to
the cell receptor protein to initiate infection. This
is true in the case of HIV.
Accordingly, the present invention provides for a
method of treating a viral (or bacterial, protozoan,
mycoplasmal, or fungal) infection comprising adminis-
tering, to a subject in need of such treatment, an
effective amount of a composition comprising ICLP
conjugate. The term "treating" as used herein refers
to an amelioration in the clinical condition of the
subject, and does not necessarily indicate that a
complete cure has been achieved. An amelioration in
clinical condition refers to a prolonged survival, a
decreased duration of illness, or a subjective
improvement in the quality of life of the subject.
The present invention provides for a method of
enhancing an immune response directed toward a viral,
protozoan, mycoplasmal, bacterial or fungal pathogen,
in a subject in need of such treatment, comprising
administering, to a subject in need of such treatment,
an effective amount of a composition comprising ICLP
conjugate. The phrase "enhancing an immune response"
refers to an increase in cellular and/or humoral
immunity. In preferred embodiments, the amount of
cellular and/or humoral immunity is increased in the
subject by at least 25 percent. Such an enhanced
immune response may be desirable during the course of
infection, or before infection may have occurred (for
example, in the context of a vaccine).
The present invention also provides for a method
of treating a malignancy or other neoplasm comprising
a~ ; n; ~tering, to a subject in need of such treatment,
an effective amount of a composition comprising ICLP
conjugate. For example, such a method may utilize an

CA 02227326 l998-Ol-l~

WO 96/36357 PCT/US96/06756
--21--


antibody specific for a tumor associated antigen to
which may be coupled a lymphokine such as GM-CSF (which
increases susceptiblity of tumor cells to lysis by
cytotoxic T lymphocytes) or interleukin-1, or a
tumoricidal agent such as a toxin. In addition to the
definition of "treating" set forth above, tumor regres-
sion, such as a decrease in tumor mass or in the number
of metastases, of preferably at least 25 percent would
be considered "treating", as would non-progression of
disease.
Further, the present invention provides for a
method of enhancing an immune response directed toward
a malignancy or other neoplasm, in a subject in need of
such treatment, comprising administering, to a subject
in need of such treatment, an effective amount of a
composition comprising ICLP conjugate.
In still other embodiments, the present invention
may be used to down-regulate the immune response. For
example, the peptides comprised in the ICLP may be
2 0 toleragenic or may be anti-idiotype relative to an
undesirably overproduced antibody (for example, anti-
body overproduced in the context of an autoimmune or
allergic condition). An effective amount of such ICLP
conjugate may be administered to a subject in need of
2 5 such treatment.
In order that the invention described herein is
better understood, the following examples are provided.
These examples are for purposes of illustration only
and are not intended to be construed in a limiting
sense.

CA 02227326 1998-0l-l~
W096/36357 PCT~S96/06756
-22-


6. EXAMPLE: ENZYME-MEDIATED CONJUGATION OF
IMMUNOGLOBULINS TO A PR8 INFLUENZA VIRAL
EPITOPE VIA CARBOHYDRATE RESIDUES

6.1. MATERIALS AND METHODS
6.1.1. IMMUNOGLOBULINS
- Ig-HA and Ig-NP were genetically engineered as
described in Zaghouani, et al., 1993, Science 259:224
and Zaghouani et al., 1992, J. Immunol. 148:3604. Ig-
HA is a chimeric BALB/c IgG2b molecule in which the
CDR3 loop is replaced by a T cell epitope (HA110-120)
from the HA of PR8 influenza A virus. Similarly, Ig-NP
is a BALB/c IgG2b molecule in which the CDR3 loop is
replaced by the NP147-161 CTL epitope. The murine IgG1
monoclonal antibody 7.21.2 is directed toward the pl85
neu gene product from rat, and was kindly provided by
Dr. M. Greene, University of Pennsylvania; it was puri-
fied from cell culture supernatants on a Protein A-
Sepharose column. Murine IgM monoclonal antibody L.59-
3 is directed toward toposiomerase I (Muryoi et al.,
1991, Autoimmunity 9:109) and was purified from culture
supernatants on a rat anti-murine ~ chain-Sepharose
column. Human myeloma proteins IgG1, IgA and IgM were
obtained as affinity purified proteins from Sigma.

6.1.2. SY~ C PEPTIDES
HA110-120 (SFER~ ~K~; SEQ ID NO:7) corresponds
to the amino acid residues 110-120 of HA from influenza
PR8 A virus (Zaghouani et al., 1993, Science 259:224).
NP 147-161 (TYQRTRALVRTGMDP; SEQ ID NO:16) corresponds
to the amino acid residues 147-161 of NP from influenza
PR8 A virus (Zaghouani et al., 1992, J. Immunol.
48:3604). Peptides were synthesized at Biosynthesis
Inc., Lewisville, TX, and were purified by reverse
phase HPLC. The molecular mass of the peptides was
confirmed by mass spectroscopy. KLH and BSA conjugates

CA 02227326 l998-0l-l~

wo96l363s7 PCT~S96/067~6
-23-


of HA110-120 were prepared as described in Liu et al.,
1979, Biochemistry 18:690.

6.1.3. ENZYMES
Neuraminidases from Arthrobacter ureafaciens
(90.8 U/mg protein, 10 U/ml) and Clostridium per-
fringens (4.8 U/mg protein, 48 U/ml) were obtained from
Calbiochem. Galactose oxidase (690 U/mg protein, 4.5
mg/ml) was obtained from Sigma, and N-glycosidase F
(PGN ase F, 25,000 U/mg protein, 20 U/ml) was obtained
from Boehringer Mannheim. Other chemicals were pur-
chased from Sigma unless otherwise indicated.

6.1.4. PR8 INFLUENZA A VIRUS
PR8 influenza A virus was prepared from allantoic
fluid of embryonated eggs on a sucrose gradient. The
viral preparation used in the following experiments was
W -inactivated.

6.1.5. ANTIBODIES
Rabbit anti-HA110-120 antibodies were obtained by
immunization of rabbits with KLH-HA110-120 conjugate
and affinity purified on a BSA-HA110-120-Sepharose
column as described in Brumeanu et al., 1993, J.
Immunol. Methods 160:65. Rabbit anti-mouse yl and ~
chains and goat anti-rabbit antibodies were obtained
from Boehringer M~nnheim.

6.1.6. ENZYMATIC SYNTHESIS OF ICLP CONJUGATES
Neuraminidases from Arthrobacter ureafaciens and
Clostridium perfringens (50 mU, 1:1 mixture) were first
incubated overnight at 37~C with lml of phosphate buf-
fer, pH 6, containing mouse or human immunoglobulins (1
mg) and 5 mM of CaCl2. Our prel;~;n~ry experiments
using the combination of these two neuraminidases indi-
cated complete desialylation of mouse IgG. The kine-


CA 02227326 1998-01-1~
W096/36357 PCT~S96/06756
-24-


tics of the desialylation reaction were monitored by
deter~;n;ng free NANA released in solution, as des-
cribed in Warren, 1959, J. Biol. Chem. 234:1971.
Desialylated immunoglobulins were dialyzed against PBS
(pH 7) to remove free NANA, and galactose oxidase
("GAO"; 10 U), pyridine borane ("PB"; 40 mM) and
HAC110-120 synthetic peptide (100 fold molar excess)
were added, while continuously stirring, for 48 hours
at 37~C. The velocity of the enzymatic oxidation by
GAO was studied in a peroxidase/o-tolidine coupled
system by determining the increase in absorbance at
425nm resulting from the release of hydrogen peroxide
(Avigad, 1985, Arch. Biochem. Biophys. 239:531).
Stabilization of Schiff bases between peptides and
oxidized carbohydrate residues occurred upon reduction
with PB as described in Cabacungan et al., 1982, Anal.
Biochem. 124:272. After completion of the coupling
reaction, the mixture was dialyzed against PBS in bags
with 100,000 MWCO (Spectrapor), concentrated to 0.1 ml
in ultra concentrators of 100,000 MWCO (S&S) and fur-
ther analyzed.

6.1.7. WESTERN BLOT ANALYSIS
Murine IgG-carbohydrate-linked-HA, IgM-carbo-
hydrate-linked-HA and human IgG-carbohydrate-linked-HA,
IgM-carbohydrate-linked-HA and IgA-carbohydrate-linked-
HA conjugates were analyzed by SDS-PAGE under reducing
and non-reducing conditions (Wyckof~ et al., 1977,
Anal. Biochem. 78:459). The conjugates lO~g/20~1 were
electrophoresed on l0~ polyacrylamide gels for 1 hour,
at 150 volts using PROTEAN II mini-system apparatus
(BioRad). Gels were either stained with Coomassie R-
250 or electrotransferred in semidry conditions onto
PVDF membranes (0.22~, Millipore, Waters Co.) for 30
minutes at 250 mAmps. Membranes were blocked overnight
at 4~C with 5~ fat free milk (Carnation) in PBS, washed

CA 02227326 l998-0l-l~

W096/36357 PCT/US96/06756
--25--


with PBS and incubated overnight at 4~C with 1,~Lg/ml of
affinity purified rabbit anti-HA110-120 antibodies in
1% BSA--PBS--0.05g~ Tween 20. Membranes were then washed
extensively with PBS-0.05% Tween 20 and bound rabbit
anti-peptide antibodies were revealed after 2 hours of
incubation at room temperature with 2 x 105 cpm of 125I-
goat anti-rabbit antibodies in 1% BSA-PBS. For the
identification of the immunoglobulin heavy and light
chains within the murine IgG-carbohydrate-linked-HA
conjugate, samples were analyzed by SDS--PAGE under
reducing conditions. Gels were electrotransferred on
PVDF membranes and the yl and K chains were revealed
with 2 x 105 cpm of 125I-rabbit anti-mouse yl and ~c
chains antibodies. Membranes were washed for 2 hours
at room temperature with PBS-0.05% Tween 20, dried, and
exposed onto Kodak X-OMAT films, overnight at --70~C.

6.1.8. SIZE EXCLUSION CHROMATOGRAPHY
ICLP conjugates were rendered free of unconjugated
peptides using size exclusion chromatography on a
Superose--6 HR 10/30 column (Pharmacia). Briefly, ICLP
conjugates were dialyzed against PBS, concentrated to
0.1 ml using ultraconcentrators with 100,000 MWCO and
then applied onto the Superose-6 column equilibrated in
PBS at a flow rate of 0.2 ml/min. Fractions were col-
lected every minute, and the peak fractions containing
conjugates were pooled, concentrated and used for fur-
ther investigations. The chromatographic profile was
monitored at 254nm since the synthetic peptide HAC110-
120 peptide is detectable at that wavelength.

6.1.9. ESTIMATION OF THE DEGREE OF CONJUGATION
To estimate the number of HAC110-120 peptides
attached per immunoglobulin molecule, a batch of mouse
IgG--carbohydrate-linked-HA conjugate (2 mg) was pre-
pared as described above. The IgG--carbohydrate--linked-

CA 02227326 1998-Ol-l~
W096l36357 PCT~S96/067~6
-26-


HA conjugate was then extensively dialyzed against PBS
in bags of 100,000 MWCo and traces of residual uncon-
jugated peptide were removed from ICLP conjugate by
size exclusion chromatography as described. The pres-
ence of HAC110-120 peptide coupled to the oligo-
saccharide chains of the immunoglobulin was confirmed
by Western blot developed with anti-HA110-120 anti-
bodies. To estimate the number of peptides coupled per
molecule of IgG, we first cleaved the sugar-peptide
complex from the conjugate using PGNase F as described
in Tarentino et al., 1989, Methods Cell Biol. 32:111.
Briefly, a preparation of chromatographically purified
ICLP conjugate (5 ~g in 250 ~l PBS) was boiled for five
minutes in the presence of 10% mercaptoethanol and 0.1%
SDS. The solution was then cooled on ice, and incu-
bated overnight at 37~C with PGN ase F (0.04 U) and
0.5% Nonidet P40. The reaction mixture was dialyzed
against PBS in bags with 1,000 MWC0 and the amount of
HACll0-l20 peptide released from the conjugate was fur-
ther determined by IRIA (Brumeanu et al. 1993, J.
Immunol. Methods 160:65). Briefly, 96-well microtiter
plates were coated overnight at 4~C with 5 ~g/ml of
BSA-HA110-120 in O.lM carbonate buffer, pH 9.6 and
blocked for 4 hours with 3% BSA. Mixtures of 2ng of
rabbit anti-HA110-120 antibodies in 1% BSA-PBS and
several dilutions of the solution containing N-glyco-
sidase released peptides were then added into the plate
and incubated overnight at 4~C. The plates were washed
with PBS-0.05~ Tween 20 and bound rabbit anti-peptide
antibodies were revealed with 5 x 104 cpm of goat anti-
rabbit antibodies. A standard inhibition curve was con-
structed with synthetic HA110-120 peptide and the
amount of HAC110-l20 peptide detached from IgG-car-
bohydrate-linked-HA conjugate that showed 50~ inhi-
bition was estimated. Precisely, 0.9ng of HAC110-120
peptide that was found in 75 ~l of N-glycosidase

CA 02227326 l998-0l-l~

W096/36357 PCT~S96/06756
-27-


digest, was able to inhibit 50% of the rabbit anti-
HA110-120 antibody activity. To estimate the number of
HAC110-120 peptide units per molecule of Ig, the amount
of Ig corresponding to 75~1 (86.4ng) was determined by
Biuret micro assay. Since the immunoglobulin was pre-
sent in the digest solution in the reduced form and the
peptide was mostly attached to the oligosaccharide
chains on the heavy chain of immunoglobulin, we cal-
culated that 1:1 molar ratio between the heavy chain of
IgG (50kDa) and HAC110-120 peptide (1.8 kDa) should
correspond to 16.91:1. To this, the corresponding
amounts of both heavy chains of the IgG and peptide, as
found in the assays, were integrated and the estimated
number of peptides per molecule of heavy chain of IgG
was found to be 5.68. On the average, this corresponds
to 11.4 peptide units coupled per molecule of IgG.

6.1.10. T CELL ACTIVATION ASSAY
2PK3 B lymphoma cells were used as antigen pre-
senting cells (APCs). Irradiated (2,200 rads) APCs
(104) were incubated for 48 hours in round bottom 96-
well plates with graded amounts of IGPs as follows:
murine IgG or IgG-carbohydrate-linked-HA, murine IgM or
IgM-carbohydrate-linked-HA, human IgG-carbohydrate-
linked-HA, genetically engineered Ig-HA or Ig-NP,
HA110-120 and NP147-161 synthetic peptides and W -
inactivated influenza PR8 A virus. 2 x 104 HA110-120-
specific LD1-24 T hybridoma cells (Haberman et al.,
1990, J. Immunol 145, 3087) were then added. Culture
supernatants were harvested and incubated for another
72 hours with 1.5 x 104 of IL-3 dependent DA-1 cells.
IL-3 production was used in these experiments as a
measure of the activation of HA110-120 specific T
cells, as quantitated by MTT colorimetric assay
(Mosmann, 1983, J. Immunol. Methods 65:55).

CA 02227326 l998-0l-l~

W096/36357 PCT~S96/06756
-28-

6.2. RESULTS
6.2.1. ENZYMATIC COUPLING OF HAC110-120
PEPTIDES TO IMMUNOGLOBULIN
The first step in the synthesis of ICLP conjugates
consisted of enzymatic desialylation of terminal NANA
residues of immunoglobulin-bound oligosaccharides.
Although most of the oligosaccharide chains contain
terminal NANA residues linked to adjacent Gal residues
by ~,(2-6) bonds, the presence of Gal-~,(2-3)-NANA
linkages has also been reported (Kobata et al., 1989,
Ciba Found. Symp., 145(0):224). Using a (1:1) mixture
of neuraminidase from Arthrobacter ureafaciens that is
able to cleave Gal-~,(2-6)-NANA, and neuraminidase from
Clostridium perfringens, that preferentially cleaves
Gal-~,(2-3)-NANA linkages (Corfield et al., 1983,
Biochim. Biophys. Acta. 744:121), we were able to pro-
duce greater than 95% yields of mouse and human asialo-
immunogloublins.
The optimal parameters for enzymatic desialylation
by neuraminidases, oxidation of Gal residues by GAO and
reduction of the Gal-peptide bonds with PB were opti-
mized with respect to time, pH, incubation temperature,
and molar ratios. We found that a 1:100 molar ratio of
Ig/peptide leads to saturation of the Gal residues with
HAC110-120 peptides under the conditions described in
the foregoing Materials and Methods section. Western
blot analyses of the IGP conjugates developed with
either anti-HA110-120 antibodies or anti-murine yl or K
antibodies indicate that the coupling reaction occurred
pre~erentially on the heavy ~h~;n~ o~ mouse IgG1 (Fig-
ure 2).

6.2.2. PURIFICATION OF ICLP CONJUGATES
Mouse IgG-carbohydrate-linked-HA and IgM-carbo-
hydrate-linked-HA conjugates were purified by size
exclusion chromatography on a Superose-6 column. It

CA 02227326 l998-0l-l~

W096136357 PCT~S96/06756
-29-

should be noted that dialysis of the conjugates was not
~ sufficient to completely remove the unconjugated pep-
tides used in excess in the coupling reaction. Small
amounts of residual peptide eluted in the salt volume
of the column (late peak corresponding to the elution
time of 80 minutes, Figure 3). While the peak symmetry
of the native immunoglobulins was in accord with their
globular structure, the conjugates showed loss of peak
symmetry that may be related to the attachment of lin-
ear structures such as HAC110-120 peptide. This
modification was most obvious in the case of IgM-carbo-
hydrate-linked-HA conjugate, which showed a higher
degree of coupling than the IgG-carbohydrate-linked-HA
conjugate.

6.2.3. SPECIFICITY OF COUPLING HAC110-120 TO
THE SUGAR MOIETY OF IMMUNOGLOBULINS
To determine whether or not the coupling of
HAC110-120 peptide occurs on the sugar moiety of immuno-
globulin, the N-linked oligosaccharide chains of a
chromatographically purified mouse IgG-carbohydrate-
linked-HA conjugate were hydrolyzed with N-glycosidase
(PGNase F). Preparations of non-hydrolyzed and hydrol-
yzed conjugate were analyzed in parallel for the pres-
ence of HAC110-120 peptide by Western blot developed
with rabbit anti-HA110-120 antibodies. Data depicted
in Figure 4 indicate that the enzymatic detachment of
the oligosaccharide chains from a mouse IgG-carbo-
hydrate-linked-HA preparation occurred at the aspara-
gine-N-linkage with subse~uent removal of HAC110-l20
peptide. This demonstrates that coupling of the pep-
tide was specifically targeted to the sugar moiety of
immunoglobulin.

CA 02227326 l998-0l-l~
WOs6/36357 PCT~S96/06756
-30-


6.2.4. EFFICIENCY OF ENZYMATIC COUPLING OF HAC110-
120 PEPTIDES TO THE SUGAR MOIETY OF
IMMUNOGLOBULINS
For the conditions of the coupling reaction estab-
lished in our experiments, at 1:100 molar ratio betweenIg and HAC110-120 peptide the Gal residues reached
saturation. Higher ratios showed no significant
increase in the amount of peptides attached per mole-
cule of immunoglobulin. The average number of HAC110-
120 peptides coupled per molecule of mouse IgG1 at1:100 ratio was 11.5 (Figure 5).
It should be noted that incorporation into the
HA110-120 peptide of the lipophilic quadruplet amino
acid sequence AAAL corresponding to the cleavage site
of the cathepsins did not interfere with the reactivity
o~ rabbit anti-HA110-120 antibodies (Figures 2 and 5).

6.2.5. IMMUNOGENICITY OF MOUSE ICLP CONJUGATES
To study the immunogenicity of HAC110-120 peptide
coupled to the sugar moiety of Ig we measured the pro-
liferative response of the HA110-120-specific T cell
hybridoma, LD1-24, to mouse and human ICLP conjugates.
All ICLP conjugates were chromatographically purified
prior to use in this assay. Data depicted in Figure 6
show the dose effect activation of T cells induced by
various antigens containing HA110-120 epitope such as
HA110-120 synthetic peptide, W-inactivated PR8 virus,
antigenized Ig-HA and peptidized ICLPs. The
specificity of T cell activation was confirmed by con-
trols such as: native IgG and IgM, NP 147-161 syn-
thetic peptide, and genetically antigenized Ig-NP.
Dose-dependent activation of the specific T helper
cells was obtained with all HA110-120 related antigens.
At 50~ activation, mouse IgG-carbohydrate-linked-HA
conjugate was as efficient as Ig-HA ~h;me~a and 60 fold
higher than the HA110-120 synthetic peptide itself. A

CA 02227326 l998-0l-l~

W096/36357 PCT~S96/06756
-31-


similar response was obtained for the human IgG-carbo-
hydrate-linked-HA conjugate. A mouse IgM-carbohydrate-
linked-HA preparation produced twice as much activation
- as the genetically antigenized Ig-HA, and 600 times
greater activation than the HA110-120 synthetic peptide
itself.
- 6.2.6. ENZYMATIC COUPLING OF HAC110-120 PEPTIDE TO
MOUSE AND HUMAN IMMUNOGLOBULINS OF
VARIOUS ISOTYPES
To investigate the versatility of the enzymatic
coupling procedure, we attempted to peptidize various
isotypes of mouse and human immunoglobulins. Using
similar reaction conditions as for the synthesis of
mouse IgG-carbohydrate-linked-HA conjugates, we were
able to generate other mouse and human ICLP conjugates
(Figure 7). It should be noted that mouse and human
IgG-carbohydrate-linked-HA conjugates showed remarkable
homogeneity indicating a lack of cross linking between
heavy and heavy, or heavy and light chains of the
immunoglobulins. However, a certain degree of hetero-
geneity was detected among the mouse and human IgM-
carbohydrate-linked-HA and IgA-carbohydrate-linked-HA
conjugates.

6.3. DISCUSSION
Coupling of non-immunogenic components such as
haptens and synthetic peptides to protein carriers is
an important means for studying the antigenicity of
small size molecules. However, because the chemical
coupling techniques are mostly used to generate protein
conjugates, several disadvantages frequently occur when
these conjugates are used to provoke specific immune
- responses in animals. One major complication pre-
viously encountered has been the induction of an immune
response against the carrier. A second is the gener-
ation of neo-determinants introduced by the extrinsic



_

CA 02227326 l998-Ol-l~
W096/36357 PCT~S96/06756
-32-


chemical groups of the cross-linkers used to bridge the
carrier to the peptides. Although "zero-length" cross-
linkers, as compared to other homo- and heterobifunc-
tional cross-linkers, do not introduce neodeterminants
on the conjugates and preclude the risk of toxicity,
targeting chemical cross-linkers toward particular
active groups on proteins or peptides provides no
absolute specificity.
Chimeric molecules bearing short sequences of
foreign genes represent a new tool in delivering
specific epitopes to the immunocompetent cells. How-
ever, the utilities of genetically antigenized immuno-
globulins are somewhat limited by the number of pep-
tides that can be expressed per molecule of
immunoglobulin.
Based on these considerations, the present inven-
tion was developed, which uses a novel coupling method-
ology to create linkages between carbohydrate residues
contained in self immunoglobulin and viral epitopes.
The conjugates produced by this methodology showed
efficient delivery of the viral epitopes to the immuno-
competent cells. The conjugation of the viral peptide
was enzymatically targeted to carbohydrate residues of
the immunoglobulins. Although galactose residues can
be chemically oxidized with periodate (Morell et al.,
1972, Methods Enzymol. 28:205), this particular type of
conjugation may involve other non-reducing termini of
the sugar moiety and it may also damage the fine func-
tional structures on the immunoglobulin molecules.
Galactose oxidase (Malmstrom et al., 1975, in "The
Enzymes XIIB", 3rd. Ed., Boyer P., ed., Academic Press,
NY, p. 527~, was used to generate a highly reactive C-6
aldehyde derivative on immunoglobulins. The C-6 alde-
hyde group reacted selectively with the ~-amino
terminus of the peptides and formed Schiff bases that
were stabilized upon mild reduction with pyridine

CA 02227326 l998-Ol-l~

W096/36357 PCT~S96/06756
-33-


borane. The ICLP conjugates were rendered free of
unconjugated peptides by size exclusion chromatography.
Using this methodology we generated homogeneous
~ conjugates between various isotypes of mouse or human
immunoglobulins and a CD4+epitope from HA of influenza
PR8 A virus. The IgG-carbohydrate-linked-HA conjugates
showed no cross-linkage between heavy and light chains
of the immunoglobulins. The remarkable homogeneity of
these conjugates may correlate with the ability of ~
amino group of the peptide in excess to compete effi-
ciently for galactose residues with other primary
amines of the immunoglobulin, such as the ~ amino
groups of lysine residues. The ~ amino groups of
lysine residues require higher pH (29.3) to become
reactive relative to the pH used to couple the HAcIlO-
120 peptide to galactose (pH 7). However, a restrained
heterogeneity was observed in the case of IgM-carbo-
hydrate-linked-HA and IgA-carbohydrate-linked-HA conju-
gates. It is unlikely that the presence of two to
three populations of conjugates, as revealed by Western
blot analysis, may represent primarily heteropolymers
between heavy and light chains since these patterns
were not observed from any of the IgG-carbohydrate-
linked-HA conjugates. Moreover, the molecular masses
2S of the IgM-carbohydrate-linked-HA and IgA-carbohydrate-
linked-HA populations did not correspond to the mole-
cular masses of any cross-linked products between the
heavy and light chains. Molecular studies showed that
a single B cell clone may encode for the synthesis of
more than one species of glycosyltransferases, in con-
trast to the unique protein structure of immunoglobulin
(Harada et al., 1987, Anal. Biochem. 164(2):374).
Tn~ee~, a variety of bianternary and complex oligo-
saccharides were found on the same monoclonal immuno-
globulin molecules (Kobata et al., 1989, Ciba Found.Symp. 145:224). This suggests that the heterogeneity

CA 02227326 1998-01-1~

W096/36357 PCT~S96/06756
-34-


of IgM-carbohydrate-linked-HA and IgA-carbohydrate-
linked-HA conjugates may be related to the micro-
heterogeneity of the sugar moiety on these particular
immunoglobulins.
The coupling efficiency of this enzymatic method
was evaluated using the average number of HAC110-120
peptides per molecule of immunoglobulin. With mono-
clonal antibody 7.21.2, an average of 11.4 peptides
were coupled per molecule of murine IgG1 (Figure 4).
Deglycosylation of the ICLP conjugates with N-glyco-
sidase showed that most of the peptide acceptors were
located on asparagine-N-linked oligosaccharides.
Western blots of the conjugates developed with anti-
peptide, anti-yl and anti-~ antibodies revealed that
the oligosaccharide-peptide complexes were formed pre-
ferentially on the heavy chain of the immunoglobulin
molecules (Figure 2).
The coupling of peptide to the sugar moiety of
immunoglobulin enhanced significantly the immuno-
genicity of the peptide. HA110-120 peptide is recog-
nized by CD4+ T cells in association with I-Ed class II
MHC alleles (Haberman et al., 1990 J. Immunol.
145:3087). Both engineered Chi ~~iC Ig-HA and IgG-
carbohydrate-linked-HA were 40 to 60 fold more effi-
cient than HA110-120 synthetic peptide (Figure 6). It
is worth noting that IgM-carbohydrate-linked-HA was 2.5
times more efficient in stimulating T helper cells than
the genetically antigenized Ig-HA and 100 fold more
efficient than HA110-120 synthetic peptide itself.
We demonstrated that HA110-120 peptide is released
from viral HA as well as from antigenized Ig-HA within
the lysosomal compartment (Brumeanu et al., 1993,
J. Exp. Med. 178:1795). Investigations on the efficacy
in stimulating specific T helper clones with synthetic
peptides and synthetically glycosylated peptides at the
amino terminus showed no significant difference

CA 02227326 1998-01-1~

W096/36357 PCT~S96/06756
-35-


(Ishioka et al., 1992, J. Immuonol. 148:2446). Paral-
lel studies using 1H NMR spectroscopy of synthetic ~-
amino glycosylated peptides indicated that the carbo-
hydrate moiety did not change the ~-helical con-
formation of the peptide (Elofsson et al., 1993,
Carbohydrate Research 246:89.31). Indeed, the ~-
helical conformation that is required for the recog-
nition by specific T helper cells was preserved in a
synthetic N-glycosylated peptide HEL51-61 even after
association with the corresponding MHC class II allele
(Allen et al., 1987, Nature (London) 327:713). More-
over, it was shown that carbohydrates do not themselves
associate with MHC molecules and are not presented to T
helper cells (Ishioka et al., 1992, J. Immunol.
148:2446; Harding et al., 1991, Proc. Natl. Acad. Sci.
U.S.A. 88:2740). This suggests that small carbohydrate
fragments attached to the N terminus of the T helper
stimulating peptides do not exhibit down-regulating
effects on the cellular immune response. However, to
facilitate specific cleavage at the N terminus of the
HA110-120 peptide and more efficiently release the pep-
tides into lysosomal vesicles, we added a lipophilic
quadruplet AAAL (SEQ ID N0:15), that contains the
cleavage site for cathepsins to the ~-amino end of the
serine residue (Yonezawa et al., 1987, Arch. Biochem.
Biophys 256(2):499). It is known that lysosomal
cathepsins play an important role in the processing of
exogenous molecules.
Previous experiments have shown that eleven-mer
HA110-120 peptides, but not truncated HA110-117 pep-
tide, stimulated LDl-24 specific T helper cells
(Brumeanu et al., 1993, J. Exp. Med. 178:1795). Thus,
it is likely that the peptides delivered by the IGP
conjugates into the lysosomal compartment were intact
sequences of the HA110-120 peptides.

CA 02227326 l998-0l-l~
W096/36357 PCT~S96/06756
-36-


7. EXAMPLE: DEFYING THE CONVENTIONAL CLASS II
PATHWAY OF ANTIGEN PRESENTATION BY A NEW
CLASS OF ANTIGENS: IMMUNO-GLYCO-PEPTIDE
CONJUGATES
Presentation of peptides in the context of MHC
class II molecules for specific recognition by recep-
tors on CD4+ T cells has been, prior to the present
invention, acclaimed as the ultimate triggering event
in T cell activation. Dogma demanded that prior to epi-
tope presentation to T helper cells, antigen is intern-
alized by antigen processing cells, fragmented in lyso-
somes, assembled into complexes with the appropriate
MHC alleles, and finally exported to the cell surface.
Any deficiency in these events has been believed to be
likely to result in T cell unresponsiveness to the
antigen.
Although processing events were believed to neces-
sarily precede presentation, some denatured antigens,
such as ovalbumin, were shown to stimulate CD4+ T cells
prior to cell processing. This phenomenon was hypo-
thesized to result from the ability of certain T cell
subsets to recognize immunodo~;n~nt epitopes made
accessible by the denaturing process, in association
with class II antigens.
In this section, a cell-free processing pathway
for the presentation of native proteins to CD4+ T cells
is described. The pathway was identified in studies
involving enzymatically engineered Immuno-Glyco-Peptide
conjugates (IGP). The particular IGP conjugates used in
these experiments link the ~-amino terminus of the
HA110-120 immunodo ;n~nt CD4+ T cell epitope of the
hemagglutinin (HA) of influenza PR8 A virus, containing
a cleavage site for cathepsin ("HAc") to the sixth
carbon of galactose residues of immunoglobulins. The
HAc110-120 epitope was observed to elicit a specific T
helper response in association with I-Ed class II

CA 02227326 1998-01-1~

W096/36357 PCT~S96106756
-37-


alleles. The IGP conjugates were found to efficiently
activate the HA110-120 specific T cell hybridoma LD1-24
cell line in vitro.
Different indices of T cell activation were
observed for various protein carriers for HAc110-120
peptide, such as (1) HAc110-120 synthetic peptide; (2)
HA110-120 expressed in the CDR3 loop of the VH gene of
immunoglobulin (a genetically engineered Ig-HA chi-
mera); (3) HA110-120 epitope as comprised in the
bromelain released HA protein from PR8 influenza A
virus (BHA); and (4) the HAc110-120 peptide enzym-
atically conjugated to galactose residues of immuno-
globulin (mouse IgG2b-Gal-HAcllO-120 conjugate). All
these protein carriers activated the specific HA110-120
T cell hybridoma LD1-24 to different extents when pre-
sented by APC 2PK3 B cell lymphoma cells (Figure 8a).
The efficacy in stimulating LD1-24 T cells at a 50%
index of activation for both IgG2b-Gal-HAcllO-120 con-
jugate and BHA was found to be ten-fold higher than the
20 efficacy of either Ig-HA or the HAc110-120 synthetic
peptide itself. IgG2b-Gal-HAcllO-120 conjugate showed a
slightly higher T cell activation index than BHA when
administered at similar molar doses with respect to the
protein carrier.
At first, the magnitude of T cell activation
appears to depend upon the amount of epitope delivered
to APCs. Indeed, we found that Ig-HA, for example, may
charge I-Ed class II molecules with a 40-fold greater
amount of HA110-120 epitope than the synthetic peptide
itself. It would seem that the IGP conjugate, which
carries an average of four HAc110-120 peptides per
molecule of immunoglobulin, was more efective at
stimulating T cells than Ig-HA expressing two HA110-120
peptides, or BHA which expresses one peptide per pro-
tein unit.

CA 02227326 l998-Ol-l~
W096/36357 PCT~S96/06756
-38-


We further investigated the relationship between
the immunopotency of the HA110-120 epitope and various
antigen carrier moieties. We normalized the amount of
HA110-120 peptide contained in each carrier as nano-
moles of peptide per molecule of carrier. The nor-
malized amount of HA110-120 epitope was then integrated
against the index of T cell activation. In contrast to
the T cell activation indexes, the epitope immuno-
potency of BHA was slightly higher than the immuno-
potency of the IGP conjugate. Among all these proteincarriers, BHA was found to be the most potent carrier
for the HA110-120 epitope. This result may be explained
by the theory that additional co-stimulatory signals
provided by the foreign boundaries of an immunodominant
epitope encased in viral proteins may increase the
immunopotency of the epitope. Both BHA and IGP con-
jugates showed higher immunopotency than Ig-HA and the
HAc110-120 synthetic peptide itself (Figure 8b). These
results suggest that both the amount of epitope
delivered to APCs, as well as the epitope boundaries,
may be important in determining the amount of T cell
stimulation. Indeed, it has been demonstrated that
sugar moieties may endow increased immunogenicity to
the peptide determinants. For example, various sugar
polymers, such as dextrans, were successfully used as
carriers to induce elevated anti-protein antibody
titers. However, the inner m~h~n;~m(s) of triggering
efficient ; c responses by sugar carriers has
remained ambiguous.
In order to determine whether the immunogenicity
of an epitope, surrounded by a sugar moiety (such as
occur in an IGP conjugate), depends only on proteolytic
degradation in the lysosomal compartment of APCs, we
~ ~red the indexes of T cell activation obtained with
IGP conjugates carrying HA peptides with (HAc) or with-
out (HA) a cleavage site for cathepsins. We found

CA 02227326 l998-Ol-l~

wos6l36357 PCT~S96/06756
-39-


neither significant differences between the T cell
activation index of IgG2b-gal-HAcllO-120 and IgG2b-gal-
HA110-120 conjugates, nor differences in epitope
immunopotency of these two IGP conjugates.
Further investigations showed that among all car-
rier systems used in our experiments, including an IgM-
gal-HAcllO-120 conjugate, only IgG2b-gal-HA110-120 con-
jugate was able to activate LD1-24 T hybridoma cells
when they were presented by paraformaldehyde fixed
antigen presenting cells ("APCs"; Figure 9a). The
IgG2b-gal-HA110-120 conjugate was also able to activate
LD1-24 T hybridoma cells in the presence of chloro-
quine, although not to the same extent as in the
absence of chloroquine (Figure 9b). Moreover, it was
15 observed that the presentation of IgG2b-gal-HA110-120
conjugate by fixed APCs was inhibited by anti-I-Ed mAb
14-4-4S as well as rat anti-Fcy receptor mAb 24G-2
(Figure 9c). This indicates that antigen presentation
may occur in an MHC-restricted manner with additional
participation of the Fcy receptors on the surface of
APCs.
Preliminary experiments in our laboratory using
NP147-155 peptide (a CTL epitope) of nucleoprotein of
PR8 influenza virus coupled on the galactose residues
of MOPC 141 showed insignificant killing by specific
CTLs of 2PK3 used as target cells. The length of pep-
tide used in these experiments was the m; n; mum required
for class I presentation (9-10 amino acid residues).
Without being bound to any particular theory, we
hypothesize a new antigen presenting cell-free pro-
cessing model as illustrated in Figure 10. In this
model, the activation of the CD4+ T cell occurs
throughout the recognition of a MHC-peptide complex
formed in particular embodiments. The immunogenic
epitopes linked to the sugar moiety of immunoglobulin
are relatively free to assemble with class II antigens.

CA 02227326 1998-01-1~

W096/36357 PCT~S96/06756
-40-

As we found by inhibition of the T cell proliferation
assay, this interaction appears to require cell surface
stabilization of the immunoglobulin carrier by specific
receptors, such as the Fc~ receptors of APC 2PK3. Since
these experiments were performed with APCs expressing,
on their surface, a high density of class II antigens
as well as Fc receptors, one may assume that the dis-
tance between the two receptors may be important for
charging the MHC antigen with the peptide. The acces-
sibility of the conjugated peptide to interact properlywith the class II molecule may well depend on the con-
formation of carbohydrates. We found that different
immunoglobulins appear to expose the carbohydrate
moiety differently on the surface. Thus, a mouse IgG1/k
known to contain four galactose residues per molecule
of immunoglobulin was found to expose only one reactive
galactose when tested in a galactose oxidase assay.
Since the galactose residues are located in a sub-
terminal position on the carbohydrate moieties of the
immunoglobulins, one may assume that the accessibility
of these residues may play a role in the extent of
exposure of the coupled peptides on the surface of IGP
conjugates.
Various publications are cited herein which are
hereby incorporated by reference in their entireties.

CA 02227326 l998-0l-l~

WO 96l36357 PCTAUS96/06756
-41-


~Q~N~ LISTING
- (1) GENER~L INFORMATION
(i) APPLICANT: Bona, Constantin A.
Lee, Y.C.
B~ nl~ Teodor-Doru
Dehazya, Philip
(ii) TITLE OF THE lNv~hllON: CARBOHYDRATE-MEDIATED COUPLING
OF
P~ll~S TO IMMUNOGLOBULINS
(iii) NUMBER OF ~yU~N~S: 16
(iv) CORRE~uN~N~ ADDRESS:
,'A'I ADDRESSEE: Brumbaugh, Graves, Donohue & Raymond
Bl STREET: 30 Rockefeller Plaza
,C CITY: New York
I'DI STATE: NY
,EI CO~h lKY: USA
~Ft ZIP: 10112-0228
(v) COMPUTER ~n~RT.~ FORM:
(A'l MEDIUM TYPE: Di~kette
(Bl COMPUTER: IBM Compatible
(C, OPERATING SYSTEM: DOS
( D~ SOFTWARE: FastSEQ Version 1.5
(vi) ~uKR~hl APPLICATION DATA:
(A) APPLICATION NUMBER: 08/477,424
(B) FILING DATE: 07-JUNE-1995
(C) CLASSIFICATION:
(vii) PRIOR APPLICATION DATA:
(A) APPLICATION NUMBER:
(B) FILING DATE:
(viii) AllO~N~Y/AGENT INFORMATION:
(A) NAME: Clark, Richard S
(B) REGISTRATION NUMBER: 26,154
(C) K~r~K~NCE/DOCKET NUMBER: 29889-A-165/31384
(ix) TELECOMMUNICATION INFORMATION:
(A) TELEPHONE: 212-408-2558
(B) TELEFAX: 212-765-2519
(C) TELEX:
(2) INFORMATION FOR SEQ ID NO:l:
(i) ~yU~N~ CHARACTERISTICS:
~A, LENGTH: l9 amino acids
B~I TYPE: amino acid
CJ STR~NDEDNESS: ~ingle
- ~Dl TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM: Human T nn~efficiency Viru~ Type 1
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION: 301...319

CA 02227326 l998-0l-l~

W 096/36357 PCTrUS96/06756
-42-


(C) OTHER INFORMATION: Envelope Protein gpl20
(v) SEQUENCE DESCRIPTION: SEQ ID NO:l:
Arg Lys Ser Ile His Ile Gly Pro Gly Arg Ala Phe Tyr Thr Thr Gly
1 5 10 15
Glu Ile Ile

(2) INFORMATION FOR SEQ ID NO:2:
(i) SEQUENCE CHARACTERISTICS:
I'A'I LENGTH: 10 amino acids
,BI TYPE: amino acid
,'C, STRANDEDNESS: single
D TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM: Influenza Virus
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION:
(C) OTHER INFORMATION: HA1 hemagglutinin protein
(Xi) ~yU~N~ DESCRIPTION: SEQ ID NO:2:
20 Trp Leu Thr Lys Lys Gly Asp Ser Tyr Pro
1 5 10
(2) INFORMATION FOR SEQ ID NO:3:
(i) SEQUENCE CHARACTERISTICS:
~'A~ LENGTH: 10 amino acids
,'BI TYPE: amino acid
~CJ sTR~Nn~nNT..c,c ~ingle
~'DJ TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM: Influenza Virus
(iv) FEATURE:
(A) NAME/REY:
(B) LOCATION:
(C) OTHER INFORMATION: H3 protein
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:3:
Trp Leu Thr Ly~ Ser Gly Ser Thr Tyr Pro
1 5 10

(2) INFORMATION FOR SEQ ID NO:4:
( i ) ~yU~N~ CHARACTERISTICS:
,'A'I LENGTH: 10 amino acids
~B~ TYPE: amino acid
,C, STR~NDEDNESS: single
,D, TOPOLOGY: linear
( ii ) M~T~T~cuT~. TYPE: peptide

CA 02227326 1998-01-1~

W 096/36357 PCTrUS96/06756
-43-


(iii) ORIGINAL SOURCE:
(A) ORGANISM: Influenza Virus
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION:
(C) OTHER INFORMATION: H2 protein
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:4:
Trp Leu Thr Lys Glu Gly Ser Asp Tyr Pro
1 5 10
10 (2) INFORMATION FOR SEQ ID NO:5:
(i) SEQUENCE CHARACTERISTICS:
'A LENGTH: 11 amino acids
,BI TYPE: amino acid
,C, STRANDEDNESS: single
IDJ TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM: Measles Virus
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION: 404...414
(C) OTHER INFORMATION: F protein
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:5:
Ile Asn Gln Asp Pro Asp Lys Ile Leu Thr Tyr
1 5 10
(2) INFORMATION FOR SEQ ID NO:6:
(i) ~QD~N~ CHARACTERISTICS:
A'I LENGTH: 19 amino acids
IB TYPE: amino acid
,C, STR~NDEDNESS: single
~D~ TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM: Foot and Mouth Disease Virus
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION: 141...160
(C) OTHER INFORMATION: VPl protein
(xi) ~yU~N~ DESCRIPTION: SEQ ID NO:6:

40 Met Asn Ser Ala Pro Asn Leu Arg Gly Asp Leu Gln Lys Val Ala Arg
1 5 10 15
~ Thr Leu Pro

(2) INFORMATION FOR SEQ ID NO:7:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 11 amino acids

CA 02227326 l998-0l-l~

W 096/36357 PCT~US96/06756
-44-


(B) TYPE: amino acid
(C) STRANDEDNESS: single
(D) TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM: Influenza PR8A Virus
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION: 110...120
(C) OTHER INFORMATION: Hemagglutinin Protein
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:7:
Ser Phe Glu Arg Phe Glu Ile Phe Pro Lys Glu
l 5 10
(2) INFORMATION FOR SEQ ID NO:8:
(i) SEQUENCE CHA~ACTERISTICS:
,A'I LENGTH: 20 amino acids
B TYPE: amino acid
C, STRANDEDNESS: single
~Dl TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM:
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION:
(C) OTHER INFORMATION: Tetanus Toxoid Protein
(Xi) ~yU~N~ DESCRIPTION: SEQ ID NO:8:
Asn Ser Val Asp Asp Ala Leu Ile Asn Ser Thr Lys Ile Tyr Ser Tyr
l 5 10 15
Phe Pro Ser Val

(2) INFORMATION FOR SEQ ID NO:9:
( i ) ~yU~N~ CHA~ACTERISTICS:
~Al LENGTH- 17 amino acids
,BI TYPE: amino acid
~C, STRANn~nNESS: single
~D, TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM:
(iv) FEATURE:
(A) NAME/REY:
(B) LOCATION:
(C) OTHER INFORMATION: Tetanus Toxoid
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:9:

CA 02227326 l998-0l-l~

W 096/36357 PCTAUS96/06756
-45-


Pro Glu Ile Asn Gly Lys Ala Ile His Leu Val Asn Asn Glu Ser Ser
1 5 l0 15
Glu

(2) INFORMATION FOR SEQ ID NO:10:
( i ) ~yU~N~ CHARACTERISTICS:
'A' LENGTH: 15 amino acids
~B TYPE: amino acid
,C, STRANDEDNESS: single
I,D, TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM:
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION: 88...103
(C) OTHER INFORMATION: Cytochrome C Protein
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:10:
Ala Asn Glu Arg Ala Asp Leu Ile Ala Tyr Leu Gln Ala Thr Lys
1 5 10 15
(2) INFORMATION FOR SEQ ID NO:11:
( i ) ~yU~N~ CHARACTERISTICS:
'A'I LENGTH: 20 amino acids
IBI TYPE: amino acid
,,C, STRANDEDNESS: single
,D,I TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM: Mycobacteria
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION: 350...369
(C) OTHER INFORMATION: Heat Shock Protein
( Xi ) ~QU~N~ DESCRIPTION: SEQ ID NO:ll:
35 Asp Gln Val His Phe Gln Pro Leu Pro Pro Ala Val Val Lys Leu Ser
1 5 10 15
Asp Ala Leu Ile

(2) INFORMATION FOR SEQ ID NO:12:

(i) SEQUENCE CHARACTERISTICS:
~A', LENGTH: 14 amino acids
~ IBI TYPE: amino acid
,,C', STRaNDEDNESS: single
~DJ TOPOLOGY: line~r
(ii) M~T.T~'CUT.T~' TYPE: peptide
(iii) ORIGINAL SOURCE:

CA 02227326 l998-0l-l~

W 096/363S7 PCTrUS96/06756
-46-


(A) ORGANISM: Hen
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION: 48...61
(C) OTHER INFORMATION: Egg White Lysozyme
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:12:
Asp Gly Ser Thr Asp Tyr Gly Ile Leu Gln Ile Asn Ser Arg
1 5 10
(2) INFORMATION FOR SEQ ID NO:13:
(i) ~QU~N~ CHARACTERISTICS:
(A'l LENGTH: 12 amino acids
B TYPE: amino acid
~C~ STRANDEDNESS: single
~D, TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM: Streptococcus A
(iv) FEATURE:
(A) NAME/KEY:
(B).LOCATION: 308... 319
(C) OTHER INFORMATION: M Protein
(Xi) ~QUh~ DESCRIPTION: SEQ ID NO:13:
Gln Val Glu Lys Ala Leu Glu Glu Ala Asn Ser Lys
1 5 10
(2) INFORMATION FOR SEQ ID NO:14:
( i ) ~yU~N~ CHARACTERISTICS:
,'Aj LENGTH: 20 amino acids
~BJ TYPE: amino acid
,C, sTR~Nn~nN~s: single
~D~ TOPOLOGY: linear
( i i ) M~T~cuT~T! TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM: Staphylococcus sp.
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION: 81...100
(C) OTHER INFORMATION: Nucleaqe Protein
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:14:

Arg Thr Asp Lys Tyr Gly Arg Gly Leu Ala Tyr Ile Tyr Ala Asp Gly
1 5 10 15
Lys Met Val Aqn

(2) INFORMATION FOR SEQ ID NO:15:
(i) ~U~N~ CHARACTERISTICS:
(A) LENGTH: 4 amino acids

CA 02227326 1998-01-1~

W 096/36357 PCTAUS96/06756


(B) TYPE: amino acid
(C) STRANDEDNESS: single
(D) TOPOLOGY: linear
(ii) MOLECULE TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM:
(iv) FEATURE:
(A) NAME/KEY:
(B) LOCATION:
(C) OTHER INFORMATION: Synthetic Peptide
(xi) ~:Qu~N~ DESCRIPTION: SEQ ID NO:15:
Ala Ala Ala Leu

(2) INFORMATION FOR SEQ ID NO:16:
(i) SEQUENCE CHARACTERISTICS:
(Aj LENGTH: 15 amino acids
IB, TYPE: amino acid
,C, STRANDEDNESS: ~ingle
D TOPOLOGY: linear
(ii) MoT~T~'cuT~T~ TYPE: peptide
(iii) ORIGINAL SOURCE:
(A) ORGANISM: Influenza PR8A Virus
(iv) FEATURE:
(A) NAME/REY:
(B) LOCATION: 147.... 161
(C) OTHER INFORMATION: NP Protein
(xi) SEQUENCE DESCRIPTION: SEQ ID NO:16:
Thr Tyr Gln Arg Thr Arg Ala Leu Val Arg Thr Gly Met Asp Pro
1 5 10 15