Natural CD4 CD25 FOXP3 regulatory T cells in graftversus-host disease - Immune Biology of Allogeneic Hematopoietic Stem Cell Transplantation - page 253

that this may not be the case. Jones and colleagues [89] demonstrated in a
miHA-mismatch model (B10.BR → CBA) that delayed infusion of Treg cells
on day 10 after tumor inoculation and allogeneic BMT rescued recipients
from GVHD while still allowing restriction of tumor growth. Similarly, using
BLI to monitor tumor growth in allogeneic BMT recipients, we could show
that (a) Treg cells themselves are unable to mediate any GVL effect, while
(b) their cotransfer also protected tumor-bearing hosts from GVHD and
thus allowed infusion of increased Tconv cell numbers. This resulted in effi-
cient elimination of host-type leukemia and lymphoma cells from various
organs, including spleen, liver, peripheral LN, and bone marrow [84] . Simi-
lar results were obtained by others [85] , thereby supporting and confirming
the general notion that cotransfer of Treg cells, even in high numbers, does
not necessarily abrogate the concomitant GVL effect.

Since Treg cells, especially in humans, are scarce and difficult to isolate
in high purity because of the absence of a specific surface marker, many
groups have investigated the in vitro generation of iTreg from naïve CD4 + T
cells and their potency in the prevention of GVHD in animal models. Sev-
eral protocols have been established, most of them comprising activation
and expansion of CD4 + Tconv in the presence of TGF-β and/or all- trans -ret-
inoic acid and yielding Foxp3 + iTreg cells of variable stability. Consequently,
protection from GVHD upon transfer of these cell populations into alloge-
neic (or xenogeneic) hosts ranged from no or only modest to full protection
[119-122] . Clearly, more insight into iTreg differentiation and plasticity is
needed and more effective and standardized protocols for the reproducible
generation of iTreg with stable phenotype and suppressive function have to
be established before clinical use of these cells can be envisaged.

256

Treg in clinical SCT

The ability of donor Treg to prevent GVHD in murine disease models
prompted many clinical studies exploring their role in clinical SCT. The
chosen strategies included the enumeration of Treg in the graft, the moni-
toring of Treg reconstitution after SCT, and, in a few studies, the adoptive
transfer of freshly isolated or in vitro -expanded Treg from stem cell or third-
party donors.

Treg content in the graft

In GVHD models, the complete deletion of Treg from the graft accelerated
GVHD [82,83] . Thus, it seemed obvious to also correlate the Treg content of
stem cell grafts with the clinical outcome after SCT in humans. Surprisingly,
Stanzani et al. [123] initially showed that CD25 expression on CD4 and CD8
cells in peripheral blood stem cell (PBSC) grafts increased the risk for GVHD.
Yet, FOXP3 staining was not yet available then and the CD25 analysis of pre-
viously frozen samples did not permit the reliable separation of Treg from
activated T cells [123] . Follow-up studies came to other conclusions and
found decreased acute GVHD in recipients of high Treg cell numbers com-
pared to patients receiving less than the median number of Treg with their
PBSC [124-128] . Yet, Wolf and colleagues [127] observed diminished GVHD
in patients receiving high Treg cell numbers only after myeloablative, and not