**9. Auto-antigen treatment**

cause of insulin dependet diabetes [35]. A general vaccination against mumps might then either decrease the incidence of T1D, or vaccination with living virus might on the contrary initiate an autoimmune process leading to an increased incidence of T1D. None of these asso‐ ciations have been proven [36, 37]. Neither have there been any associations between vacci‐

Entero virus infections are most suspected to cause T1D. Epidemiological studies have pro‐ vided evidence of coxsackie virus (CVB) infections in subjects who later develop T1D [39]. A CVBB4 strain E2 was isolated from pancreas of a diabetic child, and the virus was then passed into islet cells and found to cause diabetes in mice, which was taken as a proof of the concept that coxsackievirus can cause T1D [40]. So far vaccination against these types of infections to

The hygiene hypothesis suggests that the immune system would deviate less often towards an autoimmune process if the immune system was occupied by an ongoing defence against serious enemies. In accordance with this hypothesis, Calmette vaccination has been tried to

Several findings indicate that the gut is involved in the development of the disease process leading to T1D [42]. The intestinal barrier may be disturbed. This might facilitate passage of proteins which could contribute to the autoimmune process. Cows milk [43], and bovine in‐ sulin in cows milk has been suggested as a possible cause of an autoimmune reaction against insulin [44]. Maturation of the immune system may also be influenced by the gut flora. Pro‐ biotics can probably influence immune function through effects on antigen-presenting cells, regulatory T cells and effector T and B cells [45] and probiotics may prevent autoimmune diabetes in NOD mice [46,47]. However, although use of probiotics would be attractive as the adverse events can be expected to be minimal, there are so far no studies proving any effect

Studies in experimental animals have shown that use of a 65-kDa heat shock protein can pre‐ vent diabetes [48]. A specific peptide, Diapep 277, seems to be the active component and this

Clinical trials in humans have shown that sc administration of Diapep 277 may preserve beta cell function in adults [49]. Thus 35 patients with type 1 diabetes and basal C-peptide above 0.1 nmol/L were assigned to subcutaneous injections of 1 mg Diapep277 and 40 mg mannitol in vegetable oil The primary endpoint was glucagon-stimulated C-peptide production. At 10 months, mean C-peptide concentrations had fallen in the placebo group (n=16) but were main‐ tained in the DiaPep277 group (n=15; p=0.039). Need for exogenous insulin was higher in the

nations against other microbes and the development of diabetes [38]

preserve beta cell function but no clinical effect has been seen [41].

**7. Heat shock protein used in immune intervention**

preserve beta cells has been disappointing,

496 Type 1 Diabetes

**6. Immune intervention by probiotics**

peptide has been tried with interesting effects.

#### **9.1. "Vaccination" with insulin**

Proinsulin and insulin and its different chains are so far the only known auto-antigens that are specific for the beta cells. Insulin has been used in trials to prevent diabetes among first degree relatives with increased risk of T1D. In Diabetes Prevention Trial-Type 1 Diabetes (DPT-1) human ultralente insulin of 0.25 units x kg/day, or placebo, was given to subjects with >50% 5-year risk of getting T1D. To give such large doses of insulin sc every day can not be regarded as immune intervention, but rather as beta cell support. In any case this type of treatment failed to reach the end-point [59].

a significantly beneficial effect of the 20μg dose of Diamyd®, and there had been very few AE,

Immune Intervention in Type I Diabetes Mellitus

http://dx.doi.org/10.5772/52801

499

A Phase IIb, randomized, double-blind, placebo-controlled multicenter Diamyd® study in 160 LADA-subjects was then performed in Sweden. Subjects received 20 μg of GAD65 or placebo on 2 occasions 4 weeks apart. The trial had a main study period of 18 months and was scheduled for unblinding in June 2007. Unfortunately, the study had to be invalidated due to concerns regarding the labeling process of the investigational product. No safety concerns were raised

To investigate safety and efficacy of Diamyd® in T1D, a Phase II clinical trial in 70 recently diagnosed T1D children and adolescents was performed [71]. The study was a randomized, double-blind, placebo-controlled multicenter study using the same dose regimen as in the successful group of the previous LADA trial. The main study period of 15 months was com‐ pleted and the trial partly unblinded for sponsor and statistician but continued blinded for all other investigators for another 15 month follow-up. Outcomes from this study provided sup‐ port for clinical safety and efficacy after administration of Diamyd®. The treatment was very well tolerated and there were no treatment-related adverse events reported still after more than 4 years follow-up. Both treatment groups showed a gradual decline from baseline of both fasting and stimulated C-peptide secretion. There was no significant effect of treatment on change in fasting C-peptide after 15 months (primary endpoint). However, there was a sig‐ nificant efficacy seen on change in fasting C-peptide after 30 months (p=0.045), which remained signficant when change in C-peptide/plasma glucose ratio was taken into account (p=0.02). Furthermore, stimulated C-peptide secretion, as measured by area under the curve (AUC), decreased significantly less in the GAD-alum treated group compared to the placebo group, both after 15 months (p=0.01) and after 30 months (p=0.04). The significant effect of treatment as change in fasting and stimulated C-peptide at month 30 remained when adjusting for du‐

However, although the c-peptide preservation was evident the insulin requirement in both treatment groups increased in the course of the study, and HbA1c, and plasma glucose levels

Duration of diabetes was very important for the efficacy of treatment (p=0.05 for fasting at month 30 and p=0.03 for stimulated C-peptide area under the curve at month 15 and 30). In patients treated within 6 months of diagnosis both fasting and stimulated C-peptide secretion (AUC), decreased significantly less in the GAD-alum treated group as compared to the placebo group over 30 months (fasting, p=0.03, and stimulated p=0.04) while no such difference was seen in patients with a longer duration of diabetes ( Fig 2). The treatment effect in the short duration was still seen after more than 4 years follow-up [72] in patients with < 6 months

duration of diabetes at treatment. There were no treatment-related adverse events.

none of them considered to be treatment related [70]

**9.3. GAD vaccination in children and adolescents**

ration of diabetes, age, gender, and baseline GADA levels.

increased during the study. HbA1c did not differ between the groups.

and no SAEs had been observed during 30 months observation.

Oral insulin is not supposed to be absorbed enough to affect blood glucose or to support re‐ maining beta cells, but such an administration can be regarded as immune intervention. The DPT-1 trial randomized 372 relatives of subjects with T1D, positive for IAA and with normal intravenous and oral glucose tolerance test (IVGTTs and OGTTs), to oral insulin 7.5 mg daily or placebo. Although the result was negative when comparing the groups with the pre-speci‐ fied inclusion criteria, subanalyses suggested that Type 1 diabetes was significantly delayed in those individuals who had higher concentrations of IAA [60]. This suggests that auto-anti‐ gen therapy may be most efficaceous in patients whose immune system reacts strongly against a certain antigen.

The first diabetes-related auto-antibodies in young children are usually IAA and therefore insulin has been tried to prevent diabetes in high risk individuals. Intranasal proinsulin had effect in experimental animals [61] but intranasal administration of insulin in high risk children had no effect [62]. Administration of the insulin B-chain can prevent diabetes in experimental animals [63]. A combination of the insulin B-chain fragment with Freunds adjuvant has been tried also in newly-diagnosed T1D adults [64]. There was effect on T-regulatory cells but no significant effect on C-peptide.

#### **9.2. GAD-vaccination**

During our studies with plasmapheresis [21] we discovered a new diabetes-related antigen, 64kD [65], which later on was found to be glutamic acid decarboxylase (GAD [66]. Auto-an‐ tibodies towards GAD are common in T1D and there are convincing results from studies of experimental animals that treatment with GAD can prevent autoimmune diabetes [67, 68]..

An adjuvanted formulation, based on Alhydrogel®, a product of Aluminum hydroxide (alum),was developed to provide a drug (Diamyd®) used for evaluation in clinical trials. Al‐ hydrogel® is used as adjuvant in vaccines for children eg DTP, Pneumococcal conjugate, Hep‐ atitis B, Hepatitis A vaccines. Aluminum salts are inducing a humoral (Th2) rather than cellular immune response. As the T1D autoimmune process is deviated towards Th1 (or cellular) re‐ sponse to autoantigens, alum is used to counteract this deviation and "steer" the response induced by GAD away towards a Th2 response. Inclusion of adjuvant is also a way to minimize the quantity of antigen required for treatment.

Diamyd® preclinical safety studies were done and caused no concerns for clinical safety. Evaluation of the effects of Diamyd® in several different animal models of autoimmune dis‐ ease did not indicate any undesirable effects on the immune system. Phase 1 studies in humans were performed 1999.A randomized, double-blind and placebo-controlled dose-finding Phase IIa study in 47 LADA demonstrated efficacy in beta cell preservation in the 20-μg group [60] There were no Serious Adverse Events (SAEs) ) and even though the number of patients was very small, this result was encouraging. Follow-up after five years completed 2008 still showed a significantly beneficial effect of the 20μg dose of Diamyd®, and there had been very few AE, none of them considered to be treatment related [70]

A Phase IIb, randomized, double-blind, placebo-controlled multicenter Diamyd® study in 160 LADA-subjects was then performed in Sweden. Subjects received 20 μg of GAD65 or placebo on 2 occasions 4 weeks apart. The trial had a main study period of 18 months and was scheduled for unblinding in June 2007. Unfortunately, the study had to be invalidated due to concerns regarding the labeling process of the investigational product. No safety concerns were raised and no SAEs had been observed during 30 months observation.

#### **9.3. GAD vaccination in children and adolescents**

5-year risk of getting T1D. To give such large doses of insulin sc every day can not be regarded as immune intervention, but rather as beta cell support. In any case this type of treatment failed

Oral insulin is not supposed to be absorbed enough to affect blood glucose or to support re‐ maining beta cells, but such an administration can be regarded as immune intervention. The DPT-1 trial randomized 372 relatives of subjects with T1D, positive for IAA and with normal intravenous and oral glucose tolerance test (IVGTTs and OGTTs), to oral insulin 7.5 mg daily or placebo. Although the result was negative when comparing the groups with the pre-speci‐ fied inclusion criteria, subanalyses suggested that Type 1 diabetes was significantly delayed in those individuals who had higher concentrations of IAA [60]. This suggests that auto-anti‐ gen therapy may be most efficaceous in patients whose immune system reacts strongly against

The first diabetes-related auto-antibodies in young children are usually IAA and therefore insulin has been tried to prevent diabetes in high risk individuals. Intranasal proinsulin had effect in experimental animals [61] but intranasal administration of insulin in high risk children had no effect [62]. Administration of the insulin B-chain can prevent diabetes in experimental animals [63]. A combination of the insulin B-chain fragment with Freunds adjuvant has been tried also in newly-diagnosed T1D adults [64]. There was effect on T-regulatory cells but no

During our studies with plasmapheresis [21] we discovered a new diabetes-related antigen, 64kD [65], which later on was found to be glutamic acid decarboxylase (GAD [66]. Auto-an‐ tibodies towards GAD are common in T1D and there are convincing results from studies of experimental animals that treatment with GAD can prevent autoimmune diabetes [67, 68]..

An adjuvanted formulation, based on Alhydrogel®, a product of Aluminum hydroxide (alum),was developed to provide a drug (Diamyd®) used for evaluation in clinical trials. Al‐ hydrogel® is used as adjuvant in vaccines for children eg DTP, Pneumococcal conjugate, Hep‐ atitis B, Hepatitis A vaccines. Aluminum salts are inducing a humoral (Th2) rather than cellular immune response. As the T1D autoimmune process is deviated towards Th1 (or cellular) re‐ sponse to autoantigens, alum is used to counteract this deviation and "steer" the response induced by GAD away towards a Th2 response. Inclusion of adjuvant is also a way to minimize

Diamyd® preclinical safety studies were done and caused no concerns for clinical safety. Evaluation of the effects of Diamyd® in several different animal models of autoimmune dis‐ ease did not indicate any undesirable effects on the immune system. Phase 1 studies in humans were performed 1999.A randomized, double-blind and placebo-controlled dose-finding Phase IIa study in 47 LADA demonstrated efficacy in beta cell preservation in the 20-μg group [60] There were no Serious Adverse Events (SAEs) ) and even though the number of patients was very small, this result was encouraging. Follow-up after five years completed 2008 still showed

to reach the end-point [59].

498 Type 1 Diabetes

a certain antigen.

significant effect on C-peptide.

the quantity of antigen required for treatment.

**9.2. GAD-vaccination**

To investigate safety and efficacy of Diamyd® in T1D, a Phase II clinical trial in 70 recently diagnosed T1D children and adolescents was performed [71]. The study was a randomized, double-blind, placebo-controlled multicenter study using the same dose regimen as in the successful group of the previous LADA trial. The main study period of 15 months was com‐ pleted and the trial partly unblinded for sponsor and statistician but continued blinded for all other investigators for another 15 month follow-up. Outcomes from this study provided sup‐ port for clinical safety and efficacy after administration of Diamyd®. The treatment was very well tolerated and there were no treatment-related adverse events reported still after more than 4 years follow-up. Both treatment groups showed a gradual decline from baseline of both fasting and stimulated C-peptide secretion. There was no significant effect of treatment on change in fasting C-peptide after 15 months (primary endpoint). However, there was a sig‐ nificant efficacy seen on change in fasting C-peptide after 30 months (p=0.045), which remained signficant when change in C-peptide/plasma glucose ratio was taken into account (p=0.02). Furthermore, stimulated C-peptide secretion, as measured by area under the curve (AUC), decreased significantly less in the GAD-alum treated group compared to the placebo group, both after 15 months (p=0.01) and after 30 months (p=0.04). The significant effect of treatment as change in fasting and stimulated C-peptide at month 30 remained when adjusting for du‐ ration of diabetes, age, gender, and baseline GADA levels.

However, although the c-peptide preservation was evident the insulin requirement in both treatment groups increased in the course of the study, and HbA1c, and plasma glucose levels increased during the study. HbA1c did not differ between the groups.

Duration of diabetes was very important for the efficacy of treatment (p=0.05 for fasting at month 30 and p=0.03 for stimulated C-peptide area under the curve at month 15 and 30). In patients treated within 6 months of diagnosis both fasting and stimulated C-peptide secretion (AUC), decreased significantly less in the GAD-alum treated group as compared to the placebo group over 30 months (fasting, p=0.03, and stimulated p=0.04) while no such difference was seen in patients with a longer duration of diabetes ( Fig 2). The treatment effect in the short duration was still seen after more than 4 years follow-up [72] in patients with < 6 months duration of diabetes at treatment. There were no treatment-related adverse events.

pos GADA. In this study the two arms of the Phase II study ( placebo resp 20 μg of GAD65 (Diamyd® ) with 30 days interval, were the same, but in addition there was a third arm where the patients got 20 μg of GAD65 (Diamyd® ) sc also at Day 90 and 270 when the patients in the other arms got placebo injections. The primary endpoint was difference in C-peptide AUC after a Mixed Meal Tolerance Test. Surprisingly the study failed! [73]. The difference in AUC was only 16-18 % between the actively treated patients and the placebo group ( p= 0.10) and the difference in fasting C-peptide was similar ( p= 0.07). However, in several prespecified subgroups the efficacy was quite pronounced (around 30-40%), and significant.. When com‐ bining Phase II and that arm in Phase III in which the patients received 2 doses of GAD-alum, then theefficacy measured both as fasting C-peptide and AUC after MMTT seems quite im‐

Immune Intervention in Type I Diabetes Mellitus

http://dx.doi.org/10.5772/52801

501

The question arises why the results in Phase III was so much weaker than in Phase II. There are some possible explanations: In Phase III the patients who received active drug by chance were more often 10-11 years old whereas patients in the placebo group more frequently were 16-20 years old than in the actively treated arms. It is well known that younger patients loose their residual insulin secretion more rapidly and therefore this difference in ages might have influenced the result. There are also other facts which may have played a role. Thus, in the Phase II trial the patients were treated in March –April and when looking at patients in Phase III who were treated in March-April there was in fact also significant effect of GAD-treatment. Finally, in the Phase II trial no vaccinations were accepted, but in Phase III Influenza-vaccina‐ tion was allowed. Unfortunately an epidemic of H1N1-flu lead to that almost all patients were vaccinated, many of them in connection with the GAD-vaccinations. In Sweden and Finland the vaccine contained squalen, suspected to influence the immune system towards auto-im‐ munity, and in these two countries there was no efficacy of GAD-treatment, while there was efficacy in other European countries. Patients in Sweden, who did not get the influenza vac‐

cination close to the GAD-treatment, had better effect of the GAD-treatment [73].

In both the Phase IIb and the European Phase III patients treated with two doses of GAD-alum got increasing GADA levels with a a maximum after 3 months and then a gradual decrease even if the concentrations of GADA remained significantly higher than in the placebo group. Four doses given in the Phase III trial lead to even higher GADA levels. Increase of GADA had neither relationship to efficacy of the vaccination, nor to adverse events. There was no change of epitopes related to development of Stiff Person Syndrome, but a rather small but significant shift in isotypes with reduced percentage of IgG1 and increased IgG3/IgG4 detected in GADalum treated patients[74], in agreement with a Th2 deviation. Spontaneous/non-stimulated and PHA-induced secretion of all cytokines was similar in samples from children receiving GAD-alum and placebo, both before and 15 months after the first injection. Cytokine secretion of IL-5, IL-10, IL-13, IL-17, IFN-γ and TNF-α, but not of IL-6 and IL-12, in response to in vitro stimulation with GAD65 increased in GAD-alum treated patients from baseline to month 15, but a continuous increase was only seen in IL-5, IL-10 and IL-13 while other cytokines remained elevated but at a stable concentration [75]. This indicates that the treatment caused a Th2-

**9.4. GAD-vaccination and the immune system**

pressive after 30 months.

Mean changes from baseline in fasting (panelA) and stimulating (Panel B) c-peptide are fiven for all patients included in intention to treat analyses in the group receiving the recombinant human 65-kD isoform of glutamic acid decarbox‐ ulase in a standard vaccine formulation with alum (GAD-alum. 35 patients) and in the group receiving placebo (34 patients). Mean changes from baseline in fasting (Panel C) and stimulated (Panel E) C-peptide levels are also shown for those patienst treated less then 6 month after receiving the diagnosis of diabetes (11 patients in GAD-alum group and 14 patients in the placebo group). Finally, mean changes from baseline in fasting (Panel D) and stimulated (Panel F) Cpeptide levels are shown for those treated 6 months or more after diagnosis (24 patients in the GAD-alum group and 20 patients in the placebo group). Stimulated C-peptide level was measured on the basis of areas under the curve in response to the mixed-meal tolerance test. I bars indicate standard errors. To convert values for C-peptide to nano‐ grams per milimeter, devided by 0.33.

**Figure 2.** Mean Changes from Baseline Levels of Fasting and Stimulating C-Peptide, According to Treatment Group and Time of Treatment Relative to Diagnosis.

The Phase II trial was followed by aPhase III trial in Europe. 334 patients age 10- 20 years were included, with diabetes duration < 3 months at screening, fasting C-peptide >0.1 nmol/l and pos GADA. In this study the two arms of the Phase II study ( placebo resp 20 μg of GAD65 (Diamyd® ) with 30 days interval, were the same, but in addition there was a third arm where the patients got 20 μg of GAD65 (Diamyd® ) sc also at Day 90 and 270 when the patients in the other arms got placebo injections. The primary endpoint was difference in C-peptide AUC after a Mixed Meal Tolerance Test. Surprisingly the study failed! [73]. The difference in AUC was only 16-18 % between the actively treated patients and the placebo group ( p= 0.10) and the difference in fasting C-peptide was similar ( p= 0.07). However, in several prespecified subgroups the efficacy was quite pronounced (around 30-40%), and significant.. When com‐ bining Phase II and that arm in Phase III in which the patients received 2 doses of GAD-alum, then theefficacy measured both as fasting C-peptide and AUC after MMTT seems quite im‐ pressive after 30 months.

The question arises why the results in Phase III was so much weaker than in Phase II. There are some possible explanations: In Phase III the patients who received active drug by chance were more often 10-11 years old whereas patients in the placebo group more frequently were 16-20 years old than in the actively treated arms. It is well known that younger patients loose their residual insulin secretion more rapidly and therefore this difference in ages might have influenced the result. There are also other facts which may have played a role. Thus, in the Phase II trial the patients were treated in March –April and when looking at patients in Phase III who were treated in March-April there was in fact also significant effect of GAD-treatment. Finally, in the Phase II trial no vaccinations were accepted, but in Phase III Influenza-vaccina‐ tion was allowed. Unfortunately an epidemic of H1N1-flu lead to that almost all patients were vaccinated, many of them in connection with the GAD-vaccinations. In Sweden and Finland the vaccine contained squalen, suspected to influence the immune system towards auto-im‐ munity, and in these two countries there was no efficacy of GAD-treatment, while there was efficacy in other European countries. Patients in Sweden, who did not get the influenza vac‐ cination close to the GAD-treatment, had better effect of the GAD-treatment [73].

#### **9.4. GAD-vaccination and the immune system**

Mean changes from baseline in fasting (panelA) and stimulating (Panel B) c-peptide are fiven for all patients included in intention to treat analyses in the group receiving the recombinant human 65-kD isoform of glutamic acid decarbox‐ ulase in a standard vaccine formulation with alum (GAD-alum. 35 patients) and in the group receiving placebo (34 patients). Mean changes from baseline in fasting (Panel C) and stimulated (Panel E) C-peptide levels are also shown for those patienst treated less then 6 month after receiving the diagnosis of diabetes (11 patients in GAD-alum group and 14 patients in the placebo group). Finally, mean changes from baseline in fasting (Panel D) and stimulated (Panel F) Cpeptide levels are shown for those treated 6 months or more after diagnosis (24 patients in the GAD-alum group and 20 patients in the placebo group). Stimulated C-peptide level was measured on the basis of areas under the curve in response to the mixed-meal tolerance test. I bars indicate standard errors. To convert values for C-peptide to nano‐

**Figure 2.** Mean Changes from Baseline Levels of Fasting and Stimulating C-Peptide, According to Treatment Group

The Phase II trial was followed by aPhase III trial in Europe. 334 patients age 10- 20 years were included, with diabetes duration < 3 months at screening, fasting C-peptide >0.1 nmol/l and

grams per milimeter, devided by 0.33.

500 Type 1 Diabetes

and Time of Treatment Relative to Diagnosis.

In both the Phase IIb and the European Phase III patients treated with two doses of GAD-alum got increasing GADA levels with a a maximum after 3 months and then a gradual decrease even if the concentrations of GADA remained significantly higher than in the placebo group. Four doses given in the Phase III trial lead to even higher GADA levels. Increase of GADA had neither relationship to efficacy of the vaccination, nor to adverse events. There was no change of epitopes related to development of Stiff Person Syndrome, but a rather small but significant shift in isotypes with reduced percentage of IgG1 and increased IgG3/IgG4 detected in GADalum treated patients[74], in agreement with a Th2 deviation. Spontaneous/non-stimulated and PHA-induced secretion of all cytokines was similar in samples from children receiving GAD-alum and placebo, both before and 15 months after the first injection. Cytokine secretion of IL-5, IL-10, IL-13, IL-17, IFN-γ and TNF-α, but not of IL-6 and IL-12, in response to in vitro stimulation with GAD65 increased in GAD-alum treated patients from baseline to month 15, but a continuous increase was only seen in IL-5, IL-10 and IL-13 while other cytokines remained elevated but at a stable concentration [75]. This indicates that the treatment caused a Th2deviation. The immunological effects were long lasting immune responses, as they remained still 48 months after the first injection [75].

(Diamyd®) or placebo subcutaneous at day 1 and 30. As the study is not powered for efficacy

Immune Intervention in Type I Diabetes Mellitus

http://dx.doi.org/10.5772/52801

503

T-cells respond to antigens presented by antigen presenting cells (APCs).DNA-vaccines can be used to present the antigen instead of delivering intact proteins. A protein encoded by a plasmid DNA can either be produced outside the APCs if the pasmidDNA is administered into a muscle, or the plasmidDNA may be taken up by the APCs where the encoded protein is presented [78]. Proteins encoded by DNA vaccines can induce different types of antigen-

Most common routes of administration are either intramuscular, which is thought to favour Th1 responses, or intradermal, which is thought to favour Th2 response. For treatment of Type 1 diabetes intradermal injection should be most interesting. Another way of skewing the re‐

Promoters from virus, eg Cytomegalovirus, can be used. Certain sequences seem to stimulate

So far DNA-vaccines to create tolerance in autoimmune disease have been tried mainly in experimental animals. Plasmid DNA encoding for proinsulin [79] as well as for the insulin B chain [80] have been used for prevention of diabetes in experimental animals. Injection of plasmidDNA encoding for GAD has been shown effective in preventing diabetes in NODmice [81], while similar effect have been seen by combining plasmidDNA encoding for a fusion protein consisting of both GAD, IgG and IL4 [82]. Treatment with a recombinant vaccinia virus expressing GAD ( rVV-GAD65) has also shown to be effective in prevention of autoimmune diabetes in NOD mice by induction of active suppression of effector T-cells [83]. IgG1 anti‐ bodies and IL-4 increased and the IgG2 was unchanged, suggesting a Th2 deviation.Before clinical use there are several problems which need to be solved. Correct dosing is necessary as wrong dose might give increased immune response and a more aggressive disease process In addition it is important to be sure that the DNA is not integrated in the host chromosome.

The traditional generally accepted view is that when a patient gets Type 1 diabetes there is no longer any capacity of the beta cells to regenerate. However, there are almost no studies on beta cell regeneration in humans. In recent years some studies suggest that the old paradigm may be wrong and that beta cells in fact can regenerate. GLP-1 might stimulate beta cell re‐ generation.and GLP-1 agonist (Exenatide) in combination with monoclonal antibodies inter‐ fering with IL-2 ( Daclizumab) was given to patients with longstanding Type 1 diabetes with

specific immune responses, and perhaps also some non-specific reactions.

Th1 response and should therefore be avoided in treatment of T1D.

Another problem might be production of antibodies against DNA.

**11. Beta cell regeneration**

sponse towards Th2 may be to co-administer plasmids encoding Th2 cytokines.

the main aim is to study safety.

**10. DNA vaccines**

As a sign of increase of T-regulatory cells we noticed an increased GAD65-induced expression of FOXP3 and TGF-β at month 15 in cells from GAD-alum treated patients compared to pla‐ cebo, and the expression of FOXP3 and TGF-β correlated positively in the GAD-alum group but not in the placebo group[77]. Still after 48 months there were clear effects on the immune system suggesting both a Th2 deviation, a decrease of activated T-cells (CD4+CD25+high) but increase of FoxP3-positive regulatory T-cells. Thus, our interpretation is that Diamyd® treat‐ ment deviated the immune system towards tolerance against the auto-antigen GAD.

#### **9.5. Other trials with GAD vaccination**

Beside the European phase III trial discussed above [73], a similar trial was started a bit later in USA (US Phase III ClinicalTrials.gov Identifier: NCT00751842 ;Jerry Palmer, PI), with the same design. The recruitment was not so fast as initially only patients >16 years old were accepted, and therefore the recruitment had just finished when the negative results of the European Phase III trial was found. This lead to that the American trial was stopped, before it can give any results. In addition another intervention trial in newly-diagnosed Type 1 diabetic patients aged 3-45 years was performed by TrialNet (TrialNet Intervention ClinicalTrials.gov Identifier: NCT00529399). Patients were randomized in a double-blind controlled study into three arms, one with subcutaneous injections of 20 μg GAD65-alum (Diamyd®) at day 1,30 and 90, a second arm with subcutaneous injections of 20 μg GAD65-alum (Diamyd®) at day 1,30 and placebo at day 90, and a third arm with placebo at all time points The study failed. No effect on C-peptide preservation was found [77]. So far little has been presented from this trial with regard to effects on the immune system. It is difficult to know what the wide age range, variation in ethnic groups, BMI etc meant for the result.

#### **9.6. Ongoing or planned GAD-alum studies**

Because of the positive results in the Swedish Phase II study and the positive results in some prespecified subgroups in the European Phase III trial, new studies are planned. As the Phase III trial failed, GAD-alum will be given as part of combination therapy, which hopefully will give a better effect on the disease process. Thus a new pilot trial is just on its way when GADalum is combined with Vitamin D, which is supposed to positively influence the dendritic cells, contribute to Th2 deviation, but also influence directly beta cell survival and insulin sensitivity. In addition a third drug, anti-inflammatory, will be given to dampen the inflam‐ mation, which might play an important and negative role beside the autoimmune process.

In addition to interventional trials at onset of Type 1 diabetes a pilot trial with the aim to prevent T1D is ongoing in southern Sweden. High risk children have been identified as part of the so called DiPiS (Diabetes Prevention in Skåne) study, in which newborn children in the general population have been screened for auto-antibodies. Children positive for GADA, plus at least one more diabetes-related autoantibody, have been treated with either 20 μg GAD65-alum (Diamyd®) or placebo subcutaneous at day 1 and 30. As the study is not powered for efficacy the main aim is to study safety.
