Introduction: acquire the DR4-DQ8 and the DR3-DQ2


 Type 1 diabetes (T1D) is an autoimmune
condition in which the insulin producing ?
cells in the pancreas are targeted and destroyed by immune cells, including
macrophages, CD4+ and CD8+ T cells, which infiltrate the
islets (1). Defective insulin production and thus hyperglycemia result from
this ? cell destruction. Severe
long-term implications can arise in T1D patients, like blindness, kidney
failure, need for lower limb amputation, and increased risk of cardiovascular
diseases (2). T1D is characterised by a strong genetic component. The main
susceptibility locus includes the HLA class II genes. 90% of children with T1D
acquire the DR4-DQ8 and the DR3-DQ2 HLA haplotypes, while the DR15-DQ6
haplotype is thought to be protective (1). Disease risk can also be affected by
other loci, including the insulin gene (INS),
the cytotoxic T-lymphocyte-associated protein 4 gene (CTLA-4), and the gene encoding the ? chain of the interleukin 2 receptor (IL2RA) (3). Environmental factors are
also involved in the disease progression. For example viral infections, and
particularly those caused by enteroviruses, are thought to promote T1D through
direct cytolysis or by triggering autoimmunity against ? cells. Moreover,
T1D-associated autoimmune responses can be triggered through molecular mimicry
due to the structural homology between ? cell and viral antigens (28). T1D has
been characterised as an epidemic and even a pandemic, while the World Health
organisation  predicted that together
with type 2 diabetes it will be the seventh leading cause of death by 2030 (4).
T1D symptoms are currently maintained by insulin injections or using an insulin
pump but there is no cure developed for the disease. That leads to a need for
the development of novel therapeutic approaches. Pancreas and islet
transplantation have been effective treatments but are limited by donor
availability and may lead to a need for life-long immunosuppression. The use of
cyclosporine showed that immunosuppression is beneficial in T1D but the severe
side-effects prevented its wide-spread use (1).

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now

 Auto-reactive T cells are considered to be the
main effectors of ?
cell destruction in T1D. ? cell antigens may initially be exposed
to the immune system due to a viral insult or local stress, resulting in
insulitis (7). The disease is characterised by three stages. Initially,
islet-specific T cells are recruited in the pancreatic lymph nodes, where they
are activated by antigen presenting cell (APC)-mediated presentation of
islet-derived self-antigens. This is followed by T cell accumulation in the
pancreatic islets, and subsequent induction of ? cell damage by Fas-Fas ligand
interaction, perforin and granzyme secretion, and pro-inflammatory cytokine
production (5). The lymphocyte infiltration in the islets of the patients was
found to be dominated by CD8+ T cells. The most prevalent CD4+
T cell subsets in the patients are TH1 and TH17,
secreting IFN-? and
IL-17, respectively (8). Several candidate autoantigens have been identified, multiple
of which are not ?
cell-specific and the autoimmune response against them is not only seen in T1D.
Anti-islet antibodies are a T1D hallmark and are used to assess the disease
progression. However, their contribution to ? cell damage is not considered to be
significant (7). The role of B cells in T1D pathogenesis remains controversial,
as they are not regarded as necessary for the disease development. In non-obese
diabetic (NOD) mice B cells were found to contribute more to the disease
progression through auto-antigen presentation and T cell activation rather than
antibody production (5). Innate cells are also detected in the islets of
patients. T1D pathogenesis is reinforced by macrophage and dendritic
cell-mediated presentation of self-antigens. More specifically, in NOD mice macrophages
demonstrate a defective ability to engulf apoptotic cells while dendritic cells
exhibite an increased capacity for T cell activation, augmenting inflammation

 Regulatory T cells (Tregs) are
important in T1D prevention. It is now accepted that the overall Foxp3+
Treg population frequency is not affected by T1D, even though some
initial reports suggested the opposite. However, functional alterations in
these cells are thought to contribute to the disease development. Susceptibility loci identified through
Genome-Wide Association Studies, included genes whose products are associated
with Treg function, like IL2,
IL2RA, IL10, and CTLA4. Tregs
derived from T1D patients were found to have reduced IL-2 sensitivity, unstable
Foxp3 expression, increased apoptosis, elevated pro-inflammatory cytokine
secretion, and an altered transcriptome (9). Overall Tregs from T1D patients
are characterised by a reduced suppressive capability, which may be accompanied
by higher effector T cell (Teff) resistance to suppression (10 and
12). Adoptive Treg transfer exhibited positive results as a T1D
therapy. The first clinical trial in children recently diagnosed with T1D, resulted
in prolonged remission and no severe adverse effects (13), while the follow-up study, a year later, displayed
the safety of repetitive administration and led to a reduced need for insulin
administration, with 2/12 children becoming insulin-independent (14). In
another study investigating the safety of such an approach no short-term
toxicities, like cytokine release syndrome or infusion reactions, were
observed. Only a few low-severity infections were detected, while no
malignancies were developed. In the same experiment the expanded Tregs
were shown to have an increased CD25, CTLA-4, and latency-associated peptide
expression, indicating an improved suppressive capacity (15).

 Treg-mediated immunosuppression
displayed positive results in the treatment of T1D. However, the outcome of
this approach is characterised by high dose-dependency as a sufficient
suppressive effect may not be induced due to low Treg dose, while a
high dose can severely immunosuppress the patient (16). Additionally, following
their administration, the ex vivo-expanded
Tregs, can migrate throughout the body. This leads to non-specific
suppression of Teff reactions and to an insufficient Treg
population in the site of autoimmunity. Thus, targeting these cells to the
pancreatic islets using a chimeric antigen receptor (CAR) could lead to an improved
effect in T1D. Tregs are also unstable in the inflammatory
environment, acquiring pro-inflammatory characteristics like IFN-? and IL-17
secretion (15). In T1D their suppressive effect in the pancreatic islets is
diminished by defects in their activation, Foxp3 expression, and survival (17).
Tumour Necrosis Factor Receptor 2 (TNFR) antagonism has been found to rescue
the impaired activation of T1D-derived Tregs in vitro (18). It is also thought to act on a costimulatory fashion
in these cells stabilizing Foxp3 expression and promoting their survival (16). Furthermore,
TNFR2+ Tregs are considered to be the most suppressive
subset (18). Thus, stimulation of this molecule would be beneficial in
generating stable site-specific Tregs with optimal suppressive
capability. “Suicide” genes, like the iCasp9 system, can also be incorporated
into the engineered Tregs, allowing control of this subset following
their injection in the organism (29).

Mouse model:

 The mouse model proposed for this experiment
is the non-obese diabetic (NOD) mouse which is the most commonly used in
diabetes research, as it exhibits significant similarity with human autoimmune
diabetes. It is a spontaneous inbred mouse model, in which insulitis is
detected 4-5 weeks after birth and through high microbial exposure it develops
into T1D. Female mice are preferably used as they tend to develop T1D earlier
in their lifetime (19). NOD mice acquire more than 30 susceptibility loci for
T1D, which closely resemble the genetic association of the disease in humans. Impaired
APC capacity to eliminate or inactivate autoreactive T cells derives from
certain MHC haplotypes encoded by some of these genes. The Tregs suppressive
capacity is also affected by these loci, mainly through the diminished IL-2
expression by Teffs and defective CTLA-4 activity. Furthermore, the
onset of T1D in NOD can be accelerated by Treg depletion, highlighting
the importance of this T cell subset (20). Genetically engineered T1D models
are also available, but due to reduced genetic and pathogenic complexity
exhibited in these mice they are mostly utilised in the investigation of
specific disease aspects (30). Lower similarity to the human disease is
detected in these mice and thus their use in the development of novel therapies
is limited. Overall, similarities in the pathogenesis of T1D in humans and NOD
mice render this model appropriate for investigating Tregs
application in T1D immunotherapy.

Treg expansion:

 In this experiment CD4+CD25+CD137-
Tregs will be isolated from mouse spleen samples by Fluorescent
Activated Cell Sorting. The CD4+CD25+Foxp3+
Tregs can then be expanded using rapamycin (22). Rapamycin is
an immunosuppressive agent that inhibits the mTOR signalling pathway, which
regulates cell growth (31). It acutely inhibits the mTOR complex 1, which is
required for the proliferation of CD4+ Teffs, including the
TH1 and TH17 subsets, but not for the growth of the Treg
population (32). Thus, it is widely used to expand both murine and human Tregs, including those
isolated from T1D patients, while depleting CD4+CD25- Teffs,
ensuring the purity of the culture (21). Anti-CD3 and anti-CD28 monoclonal
antibodies, and IL-2 are also required for the Treg expansion (22).


 Following this process, third generation CARs
will be added by reverse transcription to the cells’ genome using a lentiviral
vector. Lentiviral vectors are chosen as they acquire high transfer efficiency,
they replicate rapidly, and are safer for clinical use (24). CARs include an
extracellular antibody-derived antigen recognition domain of single-chain
Fragment variant (scFv) and an intracellular activation domain. These are
connected by a transmembrane domain. The scFv is made by multiple portions of
the heavy and light chain of a specific immunoglobulin that are fused together
using a flexible linker. This domain is connected to the transmembrane domain
by an IgG1 hinge region spacer. The transmembrane domain is a hydrophobic alpha
helix, which provides stability to the molecule. The endodomain is made by CD3?, which includes three immunoreceptor
tyrosine-based activation motifs and provides the T cell activation signal, and
CD28, which initiates the co-stimulatory signal, in order to avoid anergy and
apoptosis (24). The last molecule added to the endodomain is TNFR2, in order to
stabilize the Treg phenotype and activity (Figure 1). The
extracellular domain in this experiment will be designed to recognise the SLC30A
zinc transporter 8 protein (SLC30A8), which has been identified as a target of
autoreactive T cells in T1D (23). This protein is expressed exclusively in the
pancreatic islets of both humans and NOD mice and it is utilised in the transport
of zinc ions, which are necessary for insulin hexamer assembly (33). As a
membrane molecule it will be accessible to the transformed Tregs
enabling their activation through the CARs.

iCasp9 “suicide switch”:

 An iCasp9 “suicide system” will also be added in
the vector alongside the CAR, to allow control over the transformed Treg
population. This system was preferred over the herpes simplex virus thymidine
kinase “safety switch” as the latter is characterised by higher immunogenicity
due to its viral origin (37). The iCasp9 genetic system encodes the truncated ?Caspase9
fused to FKBP12-F36V, through a linker. The caspase recruitment domain is
removed from ?Caspase9, which only acquires its intracellular signalling domain
(34 and 39). FKBP12-F36V is a mutated FK506-binding protein, which forms
homodimers through high-affinity binding with the AP1903 small molecule
pharmaceutical (38). ?Caspase9 homodimer formation and subsequent activation is
induced by FKBP12-F36V dimerization, leading to pro-apoptotic signalling and thus
cell death (Figure 2). A truncated CD19 (?CD19)
can also be added, in order to be utilised as a selectable marker for the
transformed cells. ?CD19 consists of a
19-amino acid intracellular domain, with all the tyrosine residues that mediate
signal transduction cleaved, eliminating its biological function (35 and 42). The
genes encoding the CAR, iCasp9, and ?CD19 can be
separated using 2A-like self-cleaving sequences, which allow the simultaneous
expression of the encoded molecules (36). The T2A and P2A sequences, deriving
from the Thosea Asigna virus
and the Porcine Teschovirus-1 respectively, were chosen for this experiment, as they are thought
to acquire the higher “cleavage” capacity (40). The CAR gene will be added in
the most upstream position of the construct, which is characterised by the
highest expression, while the iCasp9 will be inserted between the two 2A-sequences
(Figure 3). Low expression of the iCasp9 fusion protein will be induced by its
position, limiting the cellular concentration of this molecule and thus
avoiding FKBP21-F36V-independent ?Caspase9
dimerization (41). Using paramagnetic microbeads
conjugated to anti-CD19 antibodies, which target the extracellular domain of ?CD19
the cells with the CAR and iCasp9-encoding regions incorporated to their genome
can be selected, further expanded, and injected back to the mouse (36).


 This method is expected to take advantage of
the beneficial effects of Treg-mediated immunosuppression, while
limiting any adverse events resulting from fully immunocompromising the
patients, like development of malignancies and increased susceptibility to
infection. The Teff responses, which are central to T1D
pathogenesis, can be diminished by Treg activity. Additionally,
APCs, including B cells can be eliminated upon Treg activation, further
repressing the autoimmune response (45). Even though T1D can be induced by
immune reactions against multiple autoantigens, the combination of which can
differ between patients, the Treg response is characterised by
“bystander suppression” (26). That term underlines that the Treg-mediated
suppression of an immune response does not require stimulation by the same
antigen that triggered the effector response (25 and 27). Hence, the Treg
effect will not be limited by the fact that SLC30A8 is a minor diabetogenic
autoantigen in NOD mice and the therapy will also be applicable in patients
with no anti-SLC30A8 autoantibodies present (33). Moreover, TNFR2 addition to
the CAR is vital for stabilizing the Treg suppressive phenotype and
thus ensuring that no pro-inflammatory properties will be acquired by these
cells following injection. Hence, through the addition of the designed CAR,
optimal pancreatic islet-specific immunosuppression will be induced by the transformed
Tregs. However, side-effects of immunosuppression may still be
evident in the pancreas. The intensity of the adverse effects can be initially
mediated by optimizing the administered dose of cells.  Upon exhibition of severe toxicity the process
can be reversed through the activation of the iCasp9 “off-switch”, which has
been clinically tested and is known to produce a rapid outcome (43). In the
event of high toxicity, liver cancer development, or infection, administration
of a single AP1903 dose can result in the elimination of up to 90% of CAR-Tregs
within 30 minutes, followed by a logarithmically increasing depletion of these
cells during the next day (36). Overall, one AP1903 dose was found to eliminate
more than 99% of cells, which acquire high transgene expression. Another major
iCasp9 advantage is that it does not derive from a separate organism, like a
virus, and thus it is characterised by low potential immunogenicity (38).

 However, there are disadvantages to this method.
Increased insensitivity of Teffs to Treg-mediated
suppression has been reported in samples derived from T1D patients (11). That
could lead to reduced effect of CAR-Treg administration and thus an
adjuvant therapy may need to be developed in order to enhance the Teff-Treg
interaction. Moreover, TNFR2 is not typically defined as a costimulatory
molecule, and even though it has exhibited that mode of action in Tregs
this will be the first approach to incorporate it on a chimeric antigen
receptor as a second costimulatory domain. There is also a downside in the use
of the iCasp9 system. Activation of this “suicide switch” leads to the induction
of apoptosis and thus its effect on the cells is irreversible (37).
Consequently, in the presence of cancer or severe liver infection the
engineered Tregs should be permanently depleted leading to reappearance
or augmentation of autoimmunity. Lastly, the use of chimeric antigen receptors
is a relatively new therapeutic approach and thus the long-term safety of these
agents has not been fully investigated yet.

 Treatment of T1D with the use of CAR-Tregs
could be combined with other therapies that are being developed. For example, site-specific
IL-2 administration has exhibited positive results in NOD mice, protecting them
from T1D development. The IL-2 pathway, which is central to Treg
function, was found to be defective in multiple T1D patients (44). Hence,
promotion of this pathway could further amplify the effect of CAR-Treg
therapy in these patients, especially following the induction of a stable Treg
phenotype with consistently high CD25 expression. The IL-2 effect is highly
dose-dependent as low-dose IL-2 has been shown to increase Treg-mediated
protection of the disease but injection of a high IL-2 dose can lead to
increased autoimmunity. This augmentation in immune system activation is
induced due to the intermediate affinity binding of this cytokine on natural
killer cells, macrophages, and T cells (44). Conclusively, the use of optimally
suppressive CAR-Tregs could restore the balance between immune
activation and suppression in type 1 diabetes and other autoimmune conditions,
while only inducing site-specific side-effects. The outcome of this therapy
could be enhanced through combination with other treatments, depending on the
underlining mechanisms of the disease in the individual, generating a more
personalised approach.