T-cell costimulation in graftversus- host disease and graft-versus-leukemia effect - Immune Biology of Allogeneic Hematopoietic Stem Cell Transplantation - page 196

evaluated. To target the CD28 receptor specifically while sparing CTLA4-
mediated negative costimulation, various anti-CD28 Abs have been tested
in the control of GVHD. Based on their stimulatory activity on T lympho-
cytes, Abs directed against CD28 can be divided into three classes: i) antag-
onistic Abs bind and block CD28 signaling, ii) agonistic Abs cross-link CD28
and prompt a costimulatory signal in synergy with TCR stimulation, iii)
superagonistic Abs induce a non-physiological CD28 engagement and a full
T-cell activation in the absence of TCR stimulation [23] .

ANTAGONISTIC MONOVALENT ANTI-CD28 ABS

Monovalent Fab fragments of anti-CD28 mAb can inhibit CD28/B7 inter-
actions without stimulating CD28. They can be used as true antagonists to
inhibit the proliferation and cytokine secretion in T lymphocytes [24] and can
induce anergy in vitro [11] . In vivo, monovalent antagonist anti-CD28 Abs
delayed acute rejection when given as monotherapy and synergized with cal-
cineurin inhibitors to prevent acute and chronic allograft rejection in kidney
and heart transplant models in nonhuman primates [21] . However, Fab frag-
ments have a very short half-life in vivo, and thus their efficacy in blocking
the CD28 receptor is limited (our unpublished observations). An alternative
approach is to create an intact chimeric Ab in which one Fab fragment is spe-
cific for CD28 and the other Fab is non-specific. Such a chimeric Ab still binds
the CD28 receptor without cross-linking, but has a long half-life in vivo [25] .

199

CONVENTIONAL BIVALENT ANTI-CD28 ABS

The action of anti-CD28 in bivalent form usually results in CD28 cross-
linking and T-cell costimulation, and the degree of cross-linking of CD28
is directly correlated with activation. CD28 transmits a molecular signal
through its association with phosphatidylinositol 3-kinase (PI3-kinase)
via the cytoplasmic domain and consequently to T-cell activation and
proliferation in conjunction with TCR stimulation. The binding of anti-
CD28 to Fcγ receptors (FcγRs) reinforces their agonist activity. Therefore,
Fc-silent anti-CD28 Abs were designed by introducing mutations into the
Fc fragment to reduce or prevent the cross-linking of CD28 through Fc/
FcγR interactions. A hamster-mouse chimeric Fc-silent anti-mouse CD28
mAb (PV1-IgG3) enabled long-term survival of heart allografts in rats by
reducing the activation of alloantigen-mediated key signaling events in
T cells [26] . FK734, a humanized Fc-silent anti-human CD28 Ab, reduced
T-cell-mediated skin allograft rejection in a humanized severe combined
immunodeficiency (SCID) model [27] and reduced epidermis thinning and
HLA-DR-positive lymphocytic infiltrates of human psoriasis plaques trans-
planted into SCID mice [28] . However, this humanized Fc-silent Ab (FK734)
still generated residual agonistic signals leading to T-cell activation and
cytokine release. In vitro, it enhanced the proliferation and interleukin-2
(IL-2) and interferon-γ (IFN-γ) secretion of CD4 + or CD8 + T lymphocytes
when stimulated with human monocytes or endothelial cells [27] , probably
as a result of the mechanical cross-linking of CD28 homodimers by this Ab.
On the other hand, in the presence of CD86-transfected monocytes, this Ab
inhibited proliferation and cytokine secretion in T lymphocytes, a phenom-
enon that could be attributed to the engagement by CD86 of the negative
costimulatory CTLA4 on responding T cells.