ava i l ab l e a t www.sc i enced i r ec t . com

C l i n i ca l Immuno logy

www.e l sev i e r . com/ loca te /yc l im

Clinical Immunology (2011) 138, 162–171

Intravenous immunoglobulins exposed to heme(heme IVIG) are more efficient than IVIGin attenuating autoimmune diabetesSladjana Pavlovic a, Nemanja Zdravkovic a, Jordan D. Dimitrovb,Aleksandar Djukic a, Nebojsa Arsenijevic a,Tchavdar L. Vassilevb, Miodrag L. Lukic a,⁎

a Center for Molecular Medicine, Faculty of Medicine, University of Kragujevac, Svetozara Markovica 69, 34000 Kragujevac, Serbiab Department of Immunology, Stephan Angeloff Institute of Microbiology, Bulgarian Academy of Sciences, Acad. G. Bonchev 26,1113 Sofia, Bulgaria

Received 14 July 2010; accepted with revision 29 October 2010Available online 30 November 2010

⁎ Corresponding author. Fax: +381 3E-mail address: m.lukic@uaeu.ac.a

1521-6616/$ – see front matter © 201doi:10.1016/j.clim.2010.10.010

KEYWORDSIVIG;Treg cells;Autoimmunity;T1D

Abstract Intravenous immunoglobulins (IVIG) are known to have a therapeutic effect in someautoimmune diseases. We examined the effect of IVIG and heme-exposed IVIG on thedevelopment of immune mediated diabetes induced in C57BL/6 mice by multiple low doses ofstreptozotocin. IVIG were used in a dose of 200 mg/kg daily for 15 days. Treatment with IVIGresulted in significant attenuation of diabetes induction as evaluated by glycemia, glycosuria andHbA1c level. Interestingly, heme-exposed IVIG had a still stronger antidiabetogenic effect.

Serum levels of proinflammatory cytokines TNF-α, IFN-γ and IL17 were lower in IVIG treatedanimals when compared with controls, while IL10 level was higher. The number of CD4+Foxp3+

cells was higher in pancreatic lymph nodes of heme-exposed IVIG treated mice. Our results showthat IVIG may downregulate diabetes induction possibly by favouring induction of T regulatorycells and suggest enhanced effect upon heme-binding to IVIG.© 2010 Elsevier Inc. All rights reserved.

4306800112.e (M.L. Lukic).

0 Elsevier Inc. All rights reser

1. Introduction

Intravenous immunoglobulins (IVIG) are a therapeutic prep-aration of normal pooled human IgG obtained from plasmapools of a large number of healthy donors that possessmultiple immunomodulatory and anti-inflammatory proper-

ved

ties [1,2]. Initially used in treating primary and secondaryimmune deficiencies, intravenous immunoglobulins areincreasingly used for the treatment of diverse autoimmuneand systemic inflammatory diseases including immunethrombocytopenia, Guillain–Barré syndrome, chronic in-flammatory demyelinating polyneuropathy, myastheniagravis, dermatomyositis, Kawasaki's syndrome, etc. [3].The mode of action of IVIG is complex and involves severalmechanisms, such as modulation of Fc receptors, interfer-ence with the complement and cytokine network, regulation

.

163Intravenous immunoglobulins exposed to heme (heme IVIG)

of cell growth, induction of apoptosis of target cells,blockade of co-stimulatory molecules, interaction with theanti-idiotypic network, neutralization of pathogenic anti-bodies and modulation of the activation and differentiationof T cells, B cells and dendritic cells [4,5].

Recently, Ravetch et al. [6] elegantly demonstrated thatthe anti-inflammatory activity of IVIG results from binding ofa minor population of IgG molecules to spleen marginal zonemacrophages expressing SIGN-1. On the other hand, Ephremet al. [7] have shown in the model of experimental allergicencephalomyelitis (EAE) that IVIG-induced protection ismediated by CD4+CD25+Foxp3+ regulatory T cells (Tregs). InNOD mice IVIG reduced the occurrence of spontaneousdiabetes but mechanism of action has not been fullyunderstood [8], although it was also suggested that benefi-cial effects of IVIG could be at least in part due to effectiveblocking of FcγRs on dendritic cells [9].

Cyclophosphamide-sensitive Tregs appear to controldiabetes development in NOD mice [10]. To that we recentlyprovided evidence that in another model of type 1 diabetes,multiple low-dose streptozotocin induced diabetes (MLD-STZ), the pathological process is also controlled by Tregs inC57BL/6 mice [11]. Heme-exposed IVIG have been shown toenhance the anti-inflammatory effects of IVIG [12].

Using the model of MLD-STZ induced diabetes weprovide here the first evidence that IVIG suppress diseaseinduction and CD4+Foxp3+ cells play a role in the immuno-modulatory effects of heme-exposed IVIG in diabetesinduction.

2. Material and methods

2.1. Experimental animals

Male 8–10 week old C57BL/6micewere used in the experiment,housed under conventional conditions and allowed laboratorychow and water ad libitum. Within each experiment, animalswere matched by age and weight (18–24 g) and randomlydivided into groups of 7–8 to receive different treatmentprotocols. The experiments were approved by the EthicsCommittee of the Faculty of Medicine in Kragujevac.

2.2. Diabetes induction

Streptozotocin (Sigma Chemical, St. Louis, MO) was dissolvedin citrate buffer of pH 4.5 and administered intraperitoneallyat a dose of 40 mg/kg/day for five consecutive days.

2.3. Native and modified IVIG

Intravenous immunoglobulin [13] was kindly provided byBulBio Ltd., Sofia, Bulgaria. The pooled IgG preparation wasexposed to heme as previously described [12]. Briefly a20 mM stock solution of heme was prepared in DMSO. Thelatter was added slowly drop wise to the immunoglobulinsolution (50 mg IgG/ml) on a magnetic stirrer to a final hemeconcentration of 10 μM. After incubation at 4 °C for 30 min,heme-exposed IVIG were passed through a sterile 0.22 nmfilter. From the first day of STZ administration mice

received s.c. injections of the native or the modified IVIGin doses of 50 mg/kg or 200 mg/kg daily for 15 days.

2.4. Cyclophosphamide administration

Cyclophosphamide (CY) was administered at a dose of175 mg/kg b.w. twice at an interval of 1-week on days 5and 7 after initiation of STZ treatment. The first CY injectionwas given approximately 8 h after the last administration ofSTZ as a precaution against a possible interaction of CY withSTZ [14].

2.5. Heme-HSA and human serum albuminadministration

Control animals received either heme-modified humanserum albumin (HSA) or HSA only . Hemin, ferriprotopor-phyrine IX chloride, (Sigma-Aldrich, St. Louis, MO) stocksolution was prepared by dissolving the dry substance in0.1 M NaOH to a final concentration of 2.5 mM. This is knownto result in transformation of hemin into hematin. In order toremove aggregates the solution was filtered through 0.22 μMfilter. The hematin solution was kept in dark at 4 °C untiluse. Human serum albumin was dissolved in saline to finalconcentration of 50 mg/ml. Finally, 800 μl of stock solution(2.5 mM) was further diluted with 100 ml HSA dissolved insaline and injected into mice (100 μl/g) in dose of 200 mg/kgdaily for 15 days. Other control group received only HSAdissolved in sterile saline and administered subcutaneouslyat a dose of 200 mg/kg per day for 15 days.

2.6. Diabetes evaluation

a) Daily blood glucose (Precision Xtra plus, Abbott, UK)measurement was performed on samples taken from the tailtip after starvation for 4 h. b) Daily urine glucose wasanalysed every day, by Uriscan 2 test strip (YD diagnostics,Korea) at 9 am. c) Glycosylated hemoglobin (HbA1c) bloodlevel was determinated on day 28 using clinical analyzer(Olympus, Center Valley, Pennsylvania).

2.7. Histological examination

Pancreata of all groups were excised and placed in 10%buffered formaldehyde fixative solution overnight at roomtemperature. For quantitative histology of infiltrating cells:10% buffered formaldehyde-fixed pancreata were routinelyembedded in paraffin wax and stained with hematoxylin andeosin. The number of infiltrating cells per islet was thenquantified in non-consecutive sections by light microscopywith a 40× oil immersion objective. To remove discrepanciesdue to variations in islet size, the mean perimetric size±standard error of mean of all islets in five non-consecutivesections from the pancreata of 3 mice per experimental andcontrol groups were computed with an Axiocam digitalcamera attached to a Zeiss Axiophot. The mean value of492.24±68.5 μmwas then used as the “standard size islet” tocalculate the number of cells per islet. Any islet larger orsmaller than the mean±SEM was excluded. All values wereexpressed as mean±SE.

164 S. Pavlovic et al.

2.8. Measurement of serum cytokines

Sera from animals were collected by single needle stick andstored at −20 °C until thawed for assay. IL17, TNF-α, IFN-γand IL10 levels were measured with highly sensitive enzyme-linked immunosorbent assay (ELISA) kits (R&D SystemsMinneapolis, MN) specific for mouse cytokines according tomanufacturer's instruction.

2.9. Flow cytometric analysis

To investigate whether the administration of immunoglobu-lins could affect the number and phenotype of lymphocytesfrom the draining lymph nodes, the numbers of lymphocyteswere measured using flow cytometric analysis (FACS).Pancreatic draining lymph nodes were removed and homog-enized in RPMI 1640 containing 5% fetal bovine serum.

2.10. Labeling of cells and intracellular staining

Isolated 0.5 million cells were washed, resuspended in coldPBS containing 0.1% sodium azide (Sigma) and then in-cubated for 10 min at 4 °C. For surface staining 0.5 millioncells were suspended in 100 μl of staining buffer (5% fetalbovine serum) containing 50 μl of FITC-anti-CD4 diluted at1:100. Cells were then incubated for 30 min at 4 °C andwashed 2× in PBS. After surface staining, cells were washedin cold PBS (flow cytometry staining buffer).

Double staining was used to detect CD4+Foxp3+ cells. Afterlabelling surface markers CD4, we conducted intracellularstaining technique for detecting Foxp3. Cells were washed incold PBS (flow cytometry staining buffer). The cell pellet wasresuspended with pulse vortex in 1 ml of freshly preparedfixation/permeability working solution.

After vortexing, cells were incubated for 2 h in thedark, washed once by adding 2 ml permeabilization bufferfollowed by centrifugation and decanting of supernatant.The washing procedure was repeated. Cells were thenincubated in 100 μl Fc block in permeabilization buffer at4 °C for 30 min in the dark. Without washing afterblocking, PE-anti-Foxp3 in 1× permeabilization buffer wasadded to the suspension and incubated for 30 min at 4 °Cin the dark. Isotype permeabilization buffer was usedinstead of fluorochrome as control. Cells were then washedin 2 ml of permeabilization buffer, centrifuged and thesupernatant decanted. This step was repeated and cellsresuspended in flow cytometry staining buffer and ana-lysed on FACSCalibur flow cytometry (Becton Dickinson,Mountain View, CA) and CellQuest software (BectonDickinson).

2.11. Statistical analysis

The unpaired Student's t-test and Mann–Whitney test wereperformed using SPSS 13.0 statistical software package (SPSSInc., Chicago, IL). Results were expressed as the means±SE.All P values were 2-sided and a P value of less than 0.05 wasconsidered statistically significant.

3. Results

3.1. High doses of IVIG and heme-exposed IVIG(200 mg/kg) attenuate diabetes induction as evaluatedby glycemia, glycosuria, HbA1c level and histology inC57BL/6 mice

3.1.1. GlycemiaMice were treated with two kinds of IVIG preparations. Theywere used in a dose of 50 mg/kg b.w. and 200 mg/kg b.w.daily for 15 consecutive days. Control animals receivedhuman serum albumin in corresponding doses. 50 mg/kg IVIGdoses did not significantly alter the development of (MLD-STZ) induced diabetes (data not shown). Mice treated withmultiple low doses of STZ and HSA became graduallyhyperglycemic. Both preparations of IVIG prevented hyper-glycemia by day 19 (Fig. 1A; pb0.05). Additionally, heme-exposed IVIG demonstrated a stronger antidiabetic effect byday 26. There was a statistical difference in glycemiabetween IVIG and heme-exposed IVIG treated group at thattime point (Fig. 1A; pb0.05).

3.1.2. GlycosuriaTreatment with higher doses of native and heme-exposedimmunoglobulins induced significant decrease of glycosuriawhich remained low until the end of the experiment whencompared with mice treated with HSA (Fig. 1B; pb0.05 byday 18). Heme-exposed IVIG treated mice did not developglycosuria to the end of the experiment (pb0.05 incomparison with native IVIG by day 18 to the end of theexperiment).

3.1.3. HbA1cHbA1c is formed by non-enzymatic glycation of free aminogroups at N-terminus of β-chain of hemoglobin A0. The levelof HbA1c is proportional to the level of glucose in bloodthroughout its erythrocytes cycle. Its measurement provesdiabetic state. Glycosylated hemoglobin (HbA1c) blood levelwas determined on day 28. We observed a positivestatistically high significant correlation between dailyaverage glycemia (R2 =0.5466; pb0.01), daily averageglycosuria (R2=0.9704; pb0.001) and HbA1c level. HbA1cblood level was significantly lower in mice treated with IVIGwhen compared to the group treated with HSA (Fig. 1C;pb0.05). However, there was no difference in the level ofHbA1c between IVIG and heme-exposed IVIG treated group.

3.1.4. Protection from insulitis in IVIG treated miceMorphological examinations of pancreata (day 28) of STZtreated mice revealed insulitis (Fig. 1D) and quantitativehistology confirmed this finding (Fig. 1E). A similar histolog-ical pattern in the islets was also found in native IVIG treatedmice, but the number of inflammatory lesions and the degreeof insulitis were decreased. In addition, mononuclearinfiltration was minimal or absent in heme-exposed IVIGtreated mice. The numbers of infiltrating cells in thepancreata of native IVIG and heme-exposed IVIG treatedmice were highly significant (pb0.001) in comparison withHSA treated group.

Figure 1 IVIG resulted in significant attenuation of diabetes induction A. As evaluated by glycemia both preparations had significantantidiabetogenic effects (pb0.05 by day 19.). However, heme-exposed IVIG was more effective and the difference from the controlreached higher level of significance at the end of the experiment (pb0.001 by the day 28, 7 mice per group). B. Native and modifiedIVIG (at 200 mg/kg) induced significant decrease of glycosuria (pb0.05 by day 18) and remained low until the end of the experimentwhen compared with HSA treated mice. Heme-exposed IVIG did not develop glycosuria by the end of the experiment (pb0.05 incomparison with native IVIG).C. HbA1c was significantly lower in mice treated with IVIG and heme-exposed IVIG (pb0.05) whencompared with controls. D. Analysis of mononuclear cell infiltration in pancreatic islets of MLD-STZ treated mice indicated mildermononuclear cell infiltration in native IVIG treated mice, while mononuclear cell infiltration in heme-exposed IVIG treated mice wasminimal or absent. E. Quantitative histology of the infiltrating cells showed highly significant (pb0.001) increased influx ofmononuclear cells in HSA treated group in comparison with IVIG treated groups. ⁎pb0.05, ⁎⁎pb0.001.

165Intravenous immunoglobulins exposed to heme (heme IVIG)

166 S. Pavlovic et al.

3.2. IVIG affect serum levels of proinflammatorycytokines after MLD-STZ diabetes induction

To study the effect of immunoglobulins treatment on secretedcytokine profiles we performed ELISA tests. TNF-α, IL17, IFN-γand IL10 levels were determined in sera of C57BL/6 mice ondays 16 and 21.We found that serum levels of proinflammatorycytokines TNF-α, IFN-γ and IL17 were higher in HSA treatedcontrols when compared with IVIG treated animals on day 16after diabetes inductionwhile IL10 levelwas lower. Treatmentof mice with IVIG preparations induced significant increase inthe serum level of IL10 (Fig. 2A;pb0.05), ondays 16 and 21anddecrease in TNF level (Fig. 2A; pb0.05) on day 16. Thedifference in the level of IL17 reached the level of statisticalsignificance on day 21. Statistically significant decreasedserum level of IFN-γ can be observed only inmice treated withheme-exposed IVIG (Fig. 2A; pb0.05; day 16).

3.3. Cyclophosphamide-sensitive CD4+Foxp3+ cellare at least partially responsible for theheme-exposed IVIG attenuation of MLD-STZinduced diabetes

It had been previously established that immunomodulatoryeffect of IVIG is at least partially mediated by CD4+Foxp3+ cells[7]. In order to get some insight into the mechanisms of theobserved effect of heme-exposed IVIGweanalysed the numberof CD4+Foxp3+ cells in pancreatic lymph nodes at the onset ofhyperglycemia and glycosuria. Treatment with low doses ofcyclophosphamide abolished antidiabetogenic effects ofheme-exposed IVIG as evaluated by glycemia (Fig. 3A) andglycosuria (Fig. 3B). As show in Fig. 3C, the numbers of Tregswere significantly higher in lymph nodes of heme-exposed IVIGtreated animals in comparison with HSA treated control(pb0.05). This increase is completely abolished by CY. Finally,cyclophosphamide treatment of heme-exposed IVIG treatedmice resulted in the lower level of IL10 in serum in comparisonwith heme-exposed IVIG treated group (pb0.05), while it wasnot significantly different from control animals treated withMLD-STZ and HSA (Fig. 3D).

3.4. Heme-HSA does not suppress MLD-STZ diabetesinduction

In order to investigate if the strong antidiabetogenic effectof heme-exposed IVIG is associated with the known anti-inflammatory effect of hematin, mice were also treated withheme-HSA only. Administered at the same dose as heme-exposed IVIG, heme-HSA did not show statistically significantantidiabetogenic effect in MLD-STZ induced disease asevaluated by glycemia (Fig. 4A), glycosuria (Fig. 4B), andthe number of CD4+Foxp3+ cells in pancreatic lymph nodes(Fig. 4C) in comparison with heme-exposed IVIG treatedgroup (pb0.05). Both HSA and hematin treated animals had asignificantly lower number of CD4+Foxp3+ cells.

4. Discussion

Intravenous immunoglobulin treatment has been reported to bebeneficial in a number of autoimmune diseases, whether

mediated by antibodies or by T cells. In the present study, wehave shown for the first time that IVIG suppress diseasedevelopment in MLD-STZ model of type 1 diabetes as evaluatedby glycemia, glycosuria, HbA1c level and histology of the islets.Mice were treated with two kinds of IVIG preparations (nativeIVIG and heme-exposed IVIG 200 mg/kg b.w. daily). Exposure ofpooled IgG to heme resulted in a considerable increase of theavailable antibody repertoire. The IgG/heme complex possessescatalytic redox-activity and acts as a potent antioxidant system[12,15]. In our study, both preparations of IVIG showedantidiabetogenic effects but heme-modified Igs were moreeffective. Earlier research showed that IVIG prevent T cellinfiltration in the CNS, thus preventing the onset of irreversibleneurologic damage and inhibiting active induction of EAE [7,16].IVIG also inhibit the development of adjuvant arthritis [17].Gong et al. noticed the same favourable therapeutic effect ofIVIG treatment in experimental autoimmune myocarditis [18].Based on the results of several controlled studies, IVIG wereproven to be effective treatments in a number of autoimmuneneurological disorders [19].

Modulation by IVIG of the production of cytokines andcytokine antagonists is a major mechanism by whichimmunoglobulin preparations exert their anti-inflammatoryeffects in vivo [20] in various inflammatory and autoimmunediseases. The anti-inflammatory effects of IVIG relating tothe modulation of cytokine production are not restricted tomonocytic cytokines, but are also largely dependent uponthe ability of IVIG to modulate T helper 1 (Th1) and T helper 2(Th2) cytokine production [21]. We have observed thedecreased production of proinflammatory cytokines TNF-αand IL17 after IVIG treatment in MLD-STZ induced diabetesmellitus. It had been shown that IVIG treatment led to adecreased production of proinflammatory cytokine TNF-α intwo experimentally induced T cell autoimmune diseases inLewis rats: experimental autoimmune encephalomyelitis andadjuvant arthritis [16,17]. Bayry et al. also demonstrated theability of IVIG to induce down-regulation of co-stimulatorymolecules associated with modulation of cytokine secretionresulting in the inhibition of autoreactive and alloreactive Tcell activation and proliferation [22]. Decreasing TNF-α serumproduction was also noticed in the model of experimentalautoimmune myocarditis in BALB/C mice [19]. Recently,Ravetch [6] demonstrated that anti-inflammatory activity ofIVIG results from binding of minor population of IgG moleculesto spleen marginal zone macrophages expressing SIGN-1. Thisengagement results in the secretion of anti-inflammatorysoluble mediators that target effector macrophages. Theseeffector macrophages respond to the anti-inflammatorymediators by increasing surface expression of the inhibitoryFcgRIIB receptor. The homologous pathway in the humandiffers in that the lectin is DC-SIGN and the regulatory cells aredendritic cells.

IVIG may exhibit its therapeutic effect, at least in part,through blocking of activating FcγRs [23] or modulatingFcγRIIB expression [24,25] on APCs and effector cells. It wasalso suggested that beneficial effects of IVIG on thespontaneous onset of T1D in the NOD lines could be at leastin part due to effective blocking of FcγRs on dendritic cells [9].

Treg cells, a central component for immune suppression,are critical for establishing self-tolerance, controllinginflammatory responses and maintaining immune homeosta-sis [26,27]. IVIG can enhance the suppressive activity of

day 16 day 21

0

100

200

300

400

500

600

700

800

pg/m

l

0

50

100

150

200

250

300

pg/m

l

TNF-α

0

100

200

300

400

500

600

700

800

pg/m

l

0

100

200

300

400

500

600

700

800

pg/m

l

IL17

0

2

4

6

8

10

12

14

16

ng/m

l

n.d.0

2

4

6

8

10

12

ng/m

l

n.d.

IL10

0

20

40

60

80

100

120

pg/m

l

0

10

20

30

40

50

60

70

80

pg/m

l

IFN-γ

saline MLD-STZ+hemeIVIG MLD-STZ+IVIG MLD-STZ+HSA

Figure 2 High doses of native and heme-exposed IVIG modifed the serum levels of proinflammatory and anti-inflammatorycytokines in C57BL/6 mice (7 mice per group). IL10 was significantly higher in IVIG and heme-exposed IVIG treated groups by day 16 and21 (pb00.05). IL17 was significantly higher by day 21 only in MLD-STZ+HSA control (pb0.05). TNF-α was significantly lower in IVIG andheme-exposed IVIG treated mice by day 16 (pb0.05). IFN-γ was significantly lower only in heme-exposed IVIG treated group whencompared with MLD-STZ+HSA control by day 16 (pb0.05). ⁎pb0.05 (n.d. — non detected).

167Intravenous immunoglobulins exposed to heme (heme IVIG)

0

2

4

6

8

10

12

14

mm

ol/l

bloo

d gl

ucos

e

0

10

20

30

40

50

mm

ol/l

urin

e gl

ucos

e

Glycemia Glycosuria

CD4+Foxp3+cells in the pancreatic lymph nodes

200mg/kg heme IVIG+CY 200mg/kg heme IVIG 200mg/kg HSA

0

2

4

6

8

10

12

14

16

ng/m

l

IL10

A B

C D

0

10

20

30

40

50

60

70

80

90

100

Mea

n x

103

± SE

M n

o.of

CD

4+ Fox

p3+

cells

per

pan

crea

tic

lym

ph n

ode

0%

2%

4%

6%

8%

10%

12%

Figure 3 By day 16th low doses of cyclophosphamide (CY) partially attenuate antidiabetogenic effect of heme-exposed IVIG asevaluated by glycemia (Fig. 3A) and glycosuria (Fig. 3B). At the same time heme-exposed IVIG treated mice had a significantly highernumber of CD4+Foxp3+ T regulatory cells when compared to other groups (Fig. 3C). Cyclophosphamide treatment led to the lower levelof IL10 in the serum in comparison with the mice treated with heme-exposed IVIG only (Fig. 3D). 8 mice per group. ⁎pb0.05.

168 S. Pavlovic et al.

CD4+CD25+ Tregs [28]. The mechanisms of IVIG-T cellinteractions are complex as IVIG may directly bind toregulatory structures on T cells, or modulate T cells functionsindirectly via soluble or cellular components of the immunesystem [29]. The mechanisms through which Tregs canfunction in vivo to block the development of type 1 diabetesare under intense investigation. Tregs present in pancreaticdraining lymph nodes regulate the priming of autoreactive T

cells by limiting their expansion and differentiation [30]. Wehave demonstrated that Tregs are implicated in the IVIG-mediated protection against diabetes mellitus (Fig. 1). Anincrease in the population of CD4+CD25+Foxp3+ Tregs isassociated with the protection against diabetes mellitus. Ourobservations explain the beneficial effects of IVIG in diverseautoimmune pathologies because Tregs are pivotal incontrolling auto-reactivity. IVIG have been shown to affect

Glycemia

heme IVIG heme HSA HSA

A B

C

0

10

20

30

40

50

60

70

80

90

100

0 11 16 21

mm

ol/l

day after STZ application

Glycosuria

0

2

4

6

8

10

12

14

16

1 9 16 21

mm

ol/l

day after STZ application

0

10

20

30

40

50

60

70

80

90

100

0%

2%

4%

6%

8%

10%

12%

Mea

n x

103

± SE

M n

o.of

CD

4+ Fox

p3+

cells

per

pan

crea

tic

lym

ph n

ode

CD4+Foxp3+cells in the pancreatic lymph nodes

Figure 4 Hematin does not significantly suppress MLD-STZ diabetes induction as evaluated by glycemia, glycosuria and number ofCD4+Foxp3+ cells in the pancreatic lymph nodes. Glycemia and glycosuria were significantly attenuated in heme-exposed IVIG whencompared with the effect of hematin treatment (pb0.05). 8 mice per group. ⁎pb0.05.

169Intravenous immunoglobulins exposed to heme (heme IVIG)

the expansion of Tregs both in vitro and in vivo [7,28]. TheIVIG-mediated protection from experimental autoimmuneencephalomyelitis (EAE) is associated with peripheral in-crease of Tregs [7]. IVIG failed to protect against EAE in micethat were depleted of Tregs population before EAE induction.Since proinflammatory cytokines can render Treg defective[31], IVIG-mediated expansion of Tregs might also be related tothe capacity of IVIG to neutralize proinflammatory cytokines[4]. However, IVIG treatment did not lead to the generation ofTregs in experimental model for myasthenia gravis [32].

As it is well known, CY causes a breakdown of regulatorynetworks by removing a suppressor cell population [33–36].Even though CY is cytotoxic to a number of lymphoidpopulations, in particular B cells [37–39], recent data hasfocused on its selective effect on CD4+CD25+ Tregs [40,41].These recent findings have pointed toward the role of CY inablating CD4+CD25+ Tregs in vivo. Brode et al. confirmed thatacceleration of type 1 diabetes by CY is associated with a

reduction in CD4+CD25+Foxp3+ Tregs [10]. In our investiga-tion low doses of cyclophosphamide partially attenuated theantidiabetogenic effect of heme-exposed IVIG which hasprompted us to believe that the effect of heme-exposed IVIGis also associated with Tregs.

It has been reported that induction of heme oxygenase-1mediates the anti-inflammatory effect of IL10 [42], and thathemin induces inhibition of nitric oxide production byastrocytes [43]. It is also suggested that heme oxygenase-1upregulation by cobalt protoporphyrin IX moderate diabetesin NOD mice [44]. However in our experiments doses of200 mg/kg heme-HSA did not induce significant antidiabeto-genic effect in MLD-STZ induced diabetes nor it affected thenumber of Tregs (Fig. 4).

Our results show for the first time that IVIG may down-regulate diabetes induction. Furthermore, heme-exposed IVIGappear to be more effective than native IVIG. This effect isaccompanied by decreasing in production of proinflamatory

170 S. Pavlovic et al.

cytokines, and was sensitive to the low doses of cyclophos-phamide and apparently dependent on CD4+Foxp3+ cells.

Acknowledgments

This work was supported by grants from the Ministry ofScience and Technological Development (project No DF145065) Belgrade and from the Faculty of Medicine,University of Kragujevac (project No JP 2/09), Serbia. Wewould like to thank Marija Zdravkovic for revising themanuscript and Dragana Djurdjevic and Ivana Milovanovicfor technical assistance.

References

[1] O. Aktas, S. Waiczies, U. Grieger, U.Wendling, R. Zschenderlein,F. Zipp, Polyspecific immunoglobulins (IVIg) suppress prolifera-tion of human (auto) antigen-specific T cells with-out inducingapoptosis, J. Neuroimmunol. 114 (2001) 160–167.

[2] M.D. Kazatchkine, S.V. Kaveri, Immunomodulation of autoim-mune and inflammatory diseases with intravenous immuneglo-bulin, N Engl J. Med. 345 (2001) 747–755.

[3] S.V. Kaveri, S. Lacroix-Desmazes, J. Bayry, The antiinflamma-tory IgG, N Engl J. Med. 359 (2008) 307–309.

[4] J. Vani, S. Elluru, V.S. Negi, S. Lacroix-Desmazes, M.D.Kazatchkine, J. Bayary, S.V. Kaveri, Role of natural antibodiesin immune homeostasis: IVIg perspective, Autoimmun. Rev. 7(2008) 440–444.

[5] S. Kaveri, N. Prasad, T. Vassilev, V. Hurez, A. Pashov, S.Lacroix-Desmazes, M. Kazatchkine, Modulation of autoimmuneresponses by intravenous immunoglobulin (IVIg), Mult. Scler. 3(1997) 121–128.

[6] R.M. Anthony, F. Wermeling, M.C. Karlsson, J.V. Ravetch,Identification of a receptor required for the anti-inflamma-tory activity of IVIG, Proc. Natl Acad. Sci. 105 (2008)19571–19578.

[7] A. Ephrem, S. Chamat, C. Miquel, S. Fisson, L. Mouthon, G.Caligiuri, S. Delignat, S. Elluru, J. Bayry, S. Lacroix-Desmazes,J.L. Cohen, B.L. Salomon, M.D. Kazatchkine, S.V. Kaveri, N.Misra, Expansion of CD4+CD25+ regulatory T cells by intravenousimmunoglobulin: a critical factor in controlling experimentalautoimmune encephalomyelitis, Blood 111 (2008) 715–722.

[8] M. Bruley-Rosset, L. Mouthon, Y. Chanseaud, F. Dhainaut, J.Lirochon, D. Bourel, Polyreactive autoantibodies purified fromhuman intravenous immunoglobulins prevent the developmentof experimental autoimmune diseases, Lab. Invest. 83 (2003)1013–1023.

[9] Y. Inoue, T. Kaifu, A. Sugahara-Tobinai, A. Nakamura, J.Miyazaki, T. Takai, Activating Fc gamma receptors participatein the development of autoimmune diabetes in NOD mice, J.Immunol. 179 (2007) 764–774.

[10] S. Brode, T. Raine, P. Zaccone, A. Cooke, Cyclophosphamide-induced type-1 diabetes in the NOD mouse is associated with areduction of CD4+CD25+Foxp3+ regulatory T cells, J. Immunol.177 (2006) 6603–6612.

[11] N. Zdravkovic, A. Shahin, N. Arsenijevic, M.L. Lukic, E.P.Mensah-Brown, Regulatory T cells and ST2 signaling controldiabetes induction with multiple low doses of streptozotocin,Mol. Immunol. 47 (2009) 28–36.

[12] J.D. Dimitrov, L.T. Roumenina, V.R. Doltchinkova, N.M.Mihaylova, S. Lacroix-Desmazes, S.V. Kaveri, T.L. Vassilev,Antibodies use hema as a cofactor to extend their pathogenelimination activity and to acquire new effector functions, J.Biol. Chem. 282 (2007) 26696–26706.

[13] I. Djoumerska, A. Tchorbanov, A. Pashov, T. Vassilev, Theautoreactivity of therapeutic intravenous immunoglobulin(IVIG) preparations depends on the fractionation methodsused, Scand. J. Immunol. 61 (2005) 357–363.

[14] V. Ablamunits, F. Quintana, T. Reshef, D. Elias, I.R. Cohen,Acceleration of autoimmune diabetes by cyclophosphamide isassociated with an enhanced IFN-gamma secretion pathway, J.Autoimmun. 13 (1999) 383–392.

[15] J.D. Dimitrov, L.T. Roumenina, V.R. Doltchinkova, T.L.Vassilev, Iron ions and haeme modulate the binding propertiesof complement subcomponent C1q and of immunoglobulins,Scand. J. Immunol. 65 (2007) 230–239.

[16] A. Achiron, F. Mor, R. Margalit, I.R. Cohen, O. Lider, S. Miron,Suppression of experimental autoimmune encephalomyelitis byintravenously administered polyclonal immunoglobulins, J.Autoimmun. 15 (2000) 323–330.

[17] A. Achiron, R. Margalit, R. Hershkoviz, D. Markovits, T. Reshef, E.Melamed, I.R. Cohen, O. Lider, Intravenous immunoglobulintreatment of experimental T cell-mediated autoimmune disease.Upregulation of T cell proliferation and downregulation of tumornecrosis factor alpha secretion, J. Clin. Invest. 93 (1994) 600–605.

[18] F. Gong, Y. Hu, L. Chen, W. Gu, The therapeutic effect ofintravenous immunoglobulins and vitamin C on the progressionof experimental autoimmune myocarditis in the mouse, Med.Sci. Monit. 13 (2007) 240–246.

[19] M. Dalakas, IVIg in other autoimmune neurological disorders:current status and future prospects, J. Neurol. 255 (2008) 12–16.

[20] U.G. Andersson, L. Bjork, U. Skansen-Saphir, J. Andersson,Pooled human IgG modulates cytokine production in lympho-cytes and monocytes, Immunol. Rev. 139 (1994) 21–42.

[21] A. Mouzaki, M. Theodoropoulou, I. Gianakopoulos, V. Vlaha, M.C.Kyrtsonis, A. Maniatis, Expression patterns of Th1 and Th2 cytokinegenes in childhood idiopathic thrombocytopenic purpura (ITP) atpresentation and their modulation by intravenous immunoglobulinG (IVIg) treatment: their role in prognosis, Blood 100 (2002)1774–1779.

[22] J. Bayry, S. Lacroix-Desmazes, C. Carbonneil, N. Misra, V.Donkova, A. Pashov, A. Chevailler, L. Mouthon, B. Weill, P.Bruneval, M.D. Kazatchkine, S.V. Kaveri, Inhibition of matura-tion and function of dendritic cells by intravenous immuno-globulin, Blood 101 (2003) 758–765.

[23] C. Ibáñez, J.B. Montoro-Ronsano, Intravenous immunoglobulinpreparations and autoimmune disorders: mechanisms of action,Curr. Pharm. Biotechnol. 4 (2003) 239–247.

[24] A. Samuelsson, T.L. Towers, J.V. Ravetch, Anti-inflammatoryactivity of IVIG mediated through the inhibitory Fc receptor,Science 291 (2001) 484–486.

[25] P. Bruhns, A. Samuelsson, J.W. Pollard, J.V. Ravetch, Colony-stimulating factor-1-dependent macrophages are responsiblefor IVIG protection in antibody-induced autoimmune disease,Immunity 18 (2003) 573–581.

[26] Mda.S. Krause, P.I. de Bittencourt Jr., Type 1 diabetes: canexercise impair the autoimmune event? The L-arginine/glutaminecoupling hypothesis, Cell Biochem. Funct. 26 (2008) 406–433.

[27] S. Sakaguchi, Regulatory T cells: key controllers of immuno-logic self-tolerance, Cell 101 (2000) 455–458.

[28] A. Kessel, H. Ammuri, R. Peri, E.R. Pavlotzky, M. Blank, Y.Shoenfeld, E. Toubi, Intravenous immunoglobulin therapyaffects T regulatory cells by increasing their suppressivefunction, J. Immunol. 179 (2007) 5571–5575.

[29] O. Aktas, F. Zipp, Regulation of self-reactive T cells by humanimmunoglobulins — implications for multiple sclerosis therapy,Curr. Pharm. Des. 9 (2003) 245–256.

[30] M.G. Roncarolo, M. Battaglia, Regulatory T-cell immuno-therapy for tolerance to self antigens and alloantigens inhumans, Nat. Rev. Immunol. 7 (2007) 585–598.

[31] J. Bayry, S. Lacroix-Desmazes, S. Dasgupta, M.D. Kazatchkine,S.V. Kaveri, Efficacy of regulatory T-cell immunotherapy: are

171Intravenous immunoglobulins exposed to heme (heme IVIG)

inflammatory cytokines key determinants? Nat. Rev. Immunol.8 (2008) 1p.

[32] S. Fuchs, T. Feferman, K.Y. Zhu, R. Meidler, R. Margalit, N.Wang, O. Laub, M.C. Souroujon, Suppression of experimentalautoimmune myasthenia gravis by intravenous immunoglobulinand isolation of a disease-specific IgG fraction, Ann. NY Acad.Sci. 1110 (2007) 550–558.

[33] L. Polak, H. Geleick, J.L. Turk, Reversal by cyclophosphamideof tolerance in contact sensitization: tolerance induced byprior feeding with DNCB, Immunology 28 (1975) 939–942.

[34] R. Yasunami, J.F. Bach, Anti-suppressor effect of cyclophos-phamide on the development of spontaneous diabetes in NODmice, Eur. J. Immunol. 18 (1988) 481–484.

[35] A. Mitsuoka, M. Baba, S. Morikawa, Enhancement of delayedhypersensitivity by depletion of suppressor T cells withcyclophosphamide in mice, Nature 262 (1976) 77–78.

[36] M. Röllinghoff, A. Starzinski-Powitz, K. Pfizenmaier, H.Wagner, Cyclophosphamide-sensitive T lymphocytes suppressthe in vivo generation ofantigen-specific cytotoxic T lympho-cytes, J. Exp. Med. 145 (1997) 455–459.

[37] B. Charlton, A. Bacelj, R.M. Slattery, T.E. Mandel, Cyclophos-phamide-induced diabetes in NOD/WEHI mice: evidence forsuppression in spontaneous autoimmune diabetes mellitus,Diabetes 38 (1989) 441–447.

[38] J.L. Turk, L.W. Poulter, Selective depletion of lymphoidtissue by cyclophosphamide, Clin. Exp. Immunol. 10 (1972)285–296.

[39] J.L. Turk, D. Parker, L.W. Poulter, Functional aspects of theselective depletion of lymphoid tissue by cyclophosphamide,Immunology 23 (1972) 493–501.

[40] M. Beyer, M. Kochanek, K. Darabi, A. Popov, M. Jensen, E. Endl,P.A. Knolle, R.K. Thomas, M. von Bergwelt-Baildon, S. Debey,M. Hallek, J.L. Schultze, Reduced frequencies and suppressivefunction of CD4+CD25 hi regulatory T cells in patients withchronic lymphocytic leukemia after therapy with fludarabine,Blood 106 (2005) 2018–2025.

[41] F. Ghiringhelli, N. Larmonier, E. Schmitt, A. Parcellier, D.Cathelin, C. Garrido, B. Chauffert, E. Solary, B. Bonnotte, F.Martin, CD4+CD25+ regulatory T cells suppress tumor immunitybut are sensitive to cyclophosphamide which allows immu-notherapy of established tumors to be curative, Eur. J.Immunol. 34 (2004) 336–344.

[42] T.S. Lee, L.Y. Chau, Heme oxygenase-1 mediates the anti-inflammatory effect of interleukin-10 in mice, Nat. Med.8 (2002) 240–246.

[43] ShengW.S. , HuS. , NettlesA.R. , LokensgardJ.R. , VercellottiG.M. ,RockR.B. , Hemin inhibits NO production by IL-1β-stimulatedhuman astrocytes through induction of heme oxygenase-1 andreduction of p38 MAPK activation, J. Neuroinflammation 7 (2010)51.

[44] M. Li, S. Peterson, D. Husney, M. Inaba, K. Guo, A. Kappas, S.Ikehara, N.G. Abraham, Long-lasting expression of HO-1 delaysprogression of type I diabetes in NOD mice, Cell Cycle 6 (2007)567–571.