Alvaro Pacheco-Silvaov, Marcus G. Bastosov, Robert A. Muggia., Oleh Pankewycz., Jean Nichols., John R. MurphyA, Terry B. Strom. and Vicki E. Rubin-Kelley. Division of Clinical Immunology., Department of Medicine, Beth Israel Hospital and Harvard Medical School, Boston, Laboratory of Immunogenetics and Transplantation+, Brigham and Women’s Hospital, The Harvard Center for the Study of Kidney Diseases, and The Department of Medicine, Harvard Medical School, Boston, Seragen., Inc., Hopkinton and Department of MedicineA, University Hospital, Boston University Medical Center, Boston

Interleukin 2 receptor targeted fusion toxin (DAB&-IL-2) treatment blocks diabetogenic autoimmunity in non-obese diabetic mice* Insulin-dependent diabetes mellitus (IDDM) is strikingly similar in the non-obese diabetic (NOD) mouse and humans. In IDDM, the systematic autoimmune destruction of insulin-producing fi cells within the pancreas is dependent on autoreactive T cells. This autoimmune process can be accelerated by transferring spleen cells from diabetic donors into irradiated syngeneic NOD mice. In a previous study we established that interleukin 2 receptor (IL 2R)-bearing cells propagated from pre-diabetic NOD mice promote IDDM. Therefore, we reasoned that specific elimination of IL 2R+ T cells should abort the diabetogenic process. Tcell expressing IL 2R can be selectively destroyed with a diphtheria toxin-related IL 2 fusion protein (DAB4g6-IL-2). We set DAB486-IL-2 the challenging task of preventing fulminant IDDM accelerated by the adoptive transfer of diabetic spleen cells. Eight weeks after the adoptive transfer only 10% and 20% of NOD mice treated with 10 and 5 pglday of DABa6-IL-2, respectively, became diabetic while 100% control mice (vehicle buffer) became diabetic within 5 weeks. A dose of 1 pg/day of DAB486-IL-2 had no protective effect. Although the protection conferred by DAB486-IL-2 is not permanent, it is maintained for at least 4 weeks following cessation of treatment. Furthermore, even though these NOD mice do eventually become diabetic, the tempo of expression and severity of diabetes, as assessed by the level of hyperglycemia, is dramatically reduced. Although histologic examination of pancreas revealed minimal degree of mononuclear infiltrate within the islets in both groups, the vehicle control mice had fewer islets per section indicating many islets had already been destroyed. In addition, spleen cells from diabetic NOD mice which were pre-treated with DAB486-IL-2 (10 pg/day) for 1 week lost their ability to transfer disease. Taken together, these studies strongly support the concept that IL 2R-bearingT cells are essential for the induction of IDDM and suggest that DAB486-IL-2 would be a promising therapeutic approach in the treatment of human IDDM.

1 Introduction

into pre-diabetic NOD mice can precipitate insulitis or diabetes [7, 81.

Insulin-dependent diabetes mellitus (IDDM) in humans is the consequence of a T cell-dependent autoimmune disease in which T cells selectively target the insulin-producing fi cells for destruction. The importance of Tcells in diabetogenic autoimmunity is emphasized by the ability of cyclosporin A (CsA) to cause remission in new onset IDDM [ 11. Studies in the non-obese diabetic mouse (NOD) indicate that it has a striking similarity to human IDDM [2]. The concept that Tcells cause IDDM in NOD mice has been further supported by recent experiments documenting that anti-Tcell mAb (anti-Thy-1.2, anti-Tcell receptor, antiCD4 and anti-CD25) prevent [3-51 or arrest [6] disease in NOD mice. Moreover, certain T cell clones captured from the islets of pre-diabetic mice with insulitis upon transfer [I 98881

*

697

IL2 fusion toxin blocks diabetes in NOD mice

Eur. J. Imrnunol. 1992. 22: 697-702

This work is supported by NIH grants DK 40839, DK 36149, P50 DK 39249, and UO1-CA 48626. Supported by Conselho Nacional de Descnvolvimento Cientifico e Tecnologico (CNPq), Brazil.

Correspondence: Vicki E. Rubin-Kelley, Laboratory of Immunogenetics and Transplantation, Brigham & Women’s Hospital, 75 Francis Street Boston, MA 0211.5, USA Abbreviations: IDDM: Insulin-dependent tus NOD: Non-obese diabetic

diabetes

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1992

mclli-

A small number of T cells become stimulated following recognition of self-Ag. These autoreactive T cells multiply and damage the targeted tissue. Obviously, interruption of autoimmune T cell activation should curtail tissue injury. Expression of the high-affinity IL 2R is a universal feature of recently activated but not resting Tcells [9, lo]. An ideal therapeutic for diabetes should rapidly and selectively destroy the activated, autoaggressive T cells. Our previous studies established that diabetogenic cells express IL 2R in vivo. In these experiments, we had used IL 2 to propagate and clone islet cell-infiltratingT cells. Adoptive transfer of a certain IL 2-dependent CD4+ IL 2R+ T cell clones into pre-diabetic NOD mice accelerated the onset of diabetes [7]. Therefore, we reasoned that specific elimination of IL 2R+ Tcells should abort the diabetogenic process. To test this hypothesis we have now treated NOD mice with the IL 2 receptor-targeted fusion toxin DAB486-IL-2. This protein is the product of a gene fusion in which the receptor-binding domain of diphtheria toxin is replaced with human IL 2 sequences [ l l ] . DAB486-IL-2 selectively binds to and is cytotoxic for only those cells which display the high-affinity IL 2R heterodimer [12]. In previous experiments we clearly established the capacity of this fusion toxin to exert potent immunosuppression in vivo by preventing delayed-type hypersensitivity (DTH) [ 131 via selective intoxication of Ag-activated T cells [ 141. Admin-

+

0014-2Y80/92/0303-0697$3..50 ,2510

698

A. Pacheco-Silva, M. G. Bastos, R. A. Muggia et al.

istration of this protein also prolongs, often indefinitely, engraftment of allogeneic heart or islet allografts [15, 161. We now show that DAB486-IL-2treatment inhibits diabetogenic autoimmunity in NOD mice. In addition, we demonstrate that diabetic NOD mice treated with DAB486IL-2 do not bear spleen cells which are capable of transferring IDDM to pre-diabetic NOD mice. Taken together, these studies support the concept that IL 2R-bearing autoimmuneT cells are essential for the induction of IDDM and that treatment of human subjects with DAB486-IL-2 is worthy of consideration.

2 Materials and methods 2.1 Animals NOD mice were bred at Brigham and Women’s Hospital. The original breeding pairs were a gift of Dr. E. Leiter (Jackson Laboratory, Bar Harbor, ME). The cumulative incidence of autoimmune diabetes mellitus in our colony reaches 95% in female and 40% in male by 6 months of age. Only female littermates were used in this study. For certain experiments, female NOD mice were purchased from Taconic Farms (Germantown, NY). All mice were kept under laminar air flow and fed a commercial diet (Purina Mills, St. Louis, MO) and rap water ad libitum. Serum samples from these mice analyzed at the Charles River Serology Laboratory (Wilmington, MA) have been consistently negative for common murine pathogens, including murine hepatitis virus, Sendai virus and mycoplasma.

Eur. J. Immunol. 1992. 22: 697-702

2.4 Treatment groups After adoptive transfer of mononuclear spleen cells, the recipient NOD mice were assigned to treatment groups. Mice were treated S.C. with 10 pglday, 5 pglday or 1 pg/day of DAB486-IL-2 or with 10 pglday of DA(197)B486-IL-2 always delivered in 0.1 ml. The vehicle controls received 0.1 ml of TBS. All treatments were initiated on the day of the adoptive transfer and continued for 4 weeks. To determine if the IL 2 component of the fusion toxin could change the tempo and severity of diabetes, mice were injected S.C. with 2.5 pglday IL 2 and compared with mice given 10 pglday DA(197)B486-IL-2or 0.1 ml of the vehicle buffer. To assess the ability of DAB486-IL-2 treatment to prevent diabetogenic autoimmunity, some diabetic NOD mice were treated with 10 pg/day DAB486-IL 2 for 1week. Spleen cells from treated or control mice were adoptively transferred to lethally irradiated 2-month-old pre-diabetic NOD mice which received no further treatment.

2.5 Disease detection Blood glucose was measured on tail vein blood, using a glucose meter (Diascan Glucometer, Home Diagnostics, Freehold, NJ), initially (weeks 1 and 2) once a week and twice a week during the period we anticipated mice would become diabetic. Hyperglycemia was defined as sustained blood glucose to > 200 mg/100 ml, i.e. 3 SD above the mean of the blood glucose level measured in pre-diabetic NOD mice.

2.2 Reagents 2.6 Histological studies DAB486-IL-2is the product of a fusion gene in which a synthetic oligonucleotide encoding a human IL 2 sequence replaces the portion of the diphtheria toxin structural gene which encodes the receptor-binding domain of the native toxin [111. DA( 197)B486-IL-2 is a mutant form of DAB486IL-2 in which a single amino acid change at position 52 (glycine to glutamic acid) results in a loss of ADP-ribosyl transferase activity. DA( 197)B48&-2 was used as a control molecule. These proteins were purified from extracts of recombinant Escherichia coli as previously described [12]. To avoid variation between differents batchs only one batch of DAB486-IL-2 that had been tested before for its immunossupressive activity was used. The fusion toxins were free of contamination by endotoxins (< 10 EUIml) and were suspended in a vehicle buffer (Tris-buffered saline; TBS), pH 7.9 and aliquoted at doses of 10,5 and 1 pg/O.l ml prior to use. Human rIL 2 was prepared in J. Murphy’s laboratory.

Mice were killed by cervical dislocation and their pancreases removed and fixed in neutral buffered formalin or snap frozen in O.C.T. compound (Miles, Inc., Elkhart, IN). Coded paraffin sections (6 pm) were stained with hematoxylin and eosin and read for the magnitude of islet lymphocytic infiltration by two individuals lacking knowledge of the treatment protocols. The degree of mononuclear cell (MC) infiltration of the islets was scored on a 0-to-3 grading scale with 0 = normal, 1 = MC surrounding islets, 2 = MC surrounding and within islets, and 3 = MC throughout islets.

2.7 Statistical analysis

x2

The test was used for comparison of the cumulative incidence of diabetes. Student’s paired t-test was used for comparison of the body weight and blood sugar levels at onset of diabetes.

2.3 Adoptive transfer of diabetes Two-month-old female NOD mice were lethally irradiated (1000 rad) and rapidly (within 4 h) injected i.v. with 1.5 x 107-2.0 x lo7 mononuclear spleen cells harvested from spontaneously diabetic NOD mice (blood glucose > 300 mg/100 ml. In our previous experiments this cell dosage uniformly induced diabetes in 2-month-old female NOD mice between 3 and 8 weeks after the cell transfer.

3 Results 3.1 DABm-IL-2 protects NOD mice from adoptively transferred diabetes DAB486-IL-2 treatment prevents diabetes in pre-diabetic NOD mice adoptively transferred with mononuclear spleen

Eur. J. Immunol. 1992. 22: 697-702

IL2 fusion toxin blocks diabetes in NOD mice

cells from diabetic NOD mice. All 8 NOD mice injected with a vehicle buffer became diabetic within 8 weeks of adoptive transfer. By comparison, only 1/9 mice injected with 10 pg/day of DABls6-IL-2 became diabetic (p < 0.001) in this period, while 4/8 of those mice injected with DA(197)B486-IL-2remained euglycemic (p < 0.007) in the same period (Fig. 1 a). Histologic examination of the group of mice receiving DAB486-IL-2 killed at 9 weeks, who remained normoglycemic after receiving diabetogenic

DA(197)8486-1L-2

----C

VEHICLE BUFFER

I

L.--&-* I

2

4

8

6

10

TREATMENT

POST ADOPTIVE TRANSFER (WK)

loo 80

0

-

I I

60-

m

e

A&

-

s g

8

I I

5.Opgd DAB486-IL-2

--A-

VEHICLE BUFER

I

I I I

40-

I

20

----C

I

-

A4

. 0 , '

.

o!

a

L .

,

12

,A. ,

16

20

TREATMENT

POST ADOPTIVE TRANSFER (WK) Figure I . Pre-diabetic NOD mice were irradiated (1000 rad) and injected i.v. with 2 X lo7 mononuclear spleen cells from diabetic NOD mice. (a) Mice were injected daily with vehicle buffer ( n = 8 ) . 10 pg of DA(197)B48h-IL-2( n = 8) or 10 pg of DAB4x6-IL-2( n =: 9) during4 weeks after adoptive transfer. (b) Mice were injected daily with vehicle buffer ( n = 8 ) , or 5 pg of DAB486-IL-2( n = 8) during 4 weeks after adoptive transfer. loot 80

6o

qc

40-

? 20

;

T

I

.

0

- -

T cells revealed minimal degree of mononuclear infiltrate (1.2 f 0.6) within the islets (11.4 f 6.7 islets/section scored). By comparison, 5/8 vehicle control-treated mice who became diabetic by 4 weeks were already dead and the 3 diabetic mice had fewer islets (6.7 f 2.0 islets/section scored), although they had similar degree of mononuclear infiltrate (0.8 k 0.8). In a second set of similar adoptive transfer experiments, 10 pg/day and 5 pg/day, but not 1 pg/day, of DAB486-IL-2 suppressed development of diabetes as compared to the vehicle-treated mice.While all (10/10) control mice (vehicle buffer) were diabetic within 5 weeks after receiving spleen cells transferred from diabetic NOD mice, only 1/10 NOD mice receiving 10 pg/day DAB486-IL-2 (p < 0.001) and 2/10 NOD mice receiving 5 pg/day DAB186-IL-2 (p