Up-regulation of fas and fasL pro-apoptotic genes expression in type 1 diabetes patients after autologous haematopoietic stem cell transplantation

Abstract

Type 1 diabetes (T1D) is a chronic autoimmune disease characterized by T cell-mediated destruction of pancreatic β cells, resulting in insulin deficiency and hyperglycaemia. Recent studies have described that apoptosis impairment during central and peripheral tolerance is involved in T1D pathogenesis. In this study, the apoptosis-related gene expression in T1D patients was evaluated before and after treatment with high-dose immunosuppression followed by autologous haematopoietic stem cell transplantation (HDI-AHSCT). We also correlated gene expression results with clinical response to HDI-AHSCT. We observed a decreased expression of bad, bax and fasL pro-apoptotic genes and an increased expression of a1, bcl-x_L and cIAP-2 anti-apoptotic genes in patients' peripheral blood mononuclear cells (PBMCs) compared to controls. After HDI-AHSCT, we found an up-regulation of fas and fasL and a down-regulation of anti-apoptotic bcl-x_L genes expression in post-HDI-AHSCT periods compared to pre-transplantation. Additionally, the levels of bad, bax, bok, fasL, bcl-x_L and cIAP-1 genes expression were found similar to controls 2 years after HDI-AHSCT. Furthermore, over-expression of pro-apoptotic noxa at 540 days post-HDI-AHSCT correlated positively with insulin-free patients and conversely with glutamic acid decarboxylase autoantibodies (GAD65) autoantibody levels. Taken together, the results suggest that apoptosis-related genes deregulation in patients' PBMCs might be involved in breakdown of immune tolerance and consequently contribute to T1D pathogenesis. Furthermore, HDI-AHSCT modulated the expression of some apoptotic genes towards the levels similar to controls. Possibly, the expression of these apoptotic molecules could be applied as biomarkers of clinical remission of T1D patients treated with HDI-AHSCT therapy.

Introduction

Type 1 diabetes (T1D) is a chronic autoimmune disease characterized by T cell-mediated destruction of pancreatic β cells, resulting in insulin deficiency and hyperglycaemia [1]–[4]. It has been estimated that approximately 78 000 children aged less than 15 years are affected annually by T1D worldwide [5]. Several factors have been described to be involved in T1D pathogenesis: immunogenetic susceptibility, represented by a strong association with histocompatibility genes, human leucocyte antigen (HLA)-DR3/DR4 [6], environmental factors [7],[8] and defects in immune regulation mechanisms, including the decrease of the suppressive activity of regulatory T cells [9],[10].

Recent studies have shown that apoptosis impairment during central and peripheral tolerance process is involved in autoimmune disease (AID) pathogenesis [11]–[13]. Apoptosis may play an important role in AIDs during lymphocyte education and selection. Thus, the apoptosis blockage contributes to the escape of autoreactive cells from immune tolerance mechanisms [14]. The central tolerance seems to depend upon the mitochondria-dependent apoptosis (intrinsic pathway), which includes the participation of the Bcl-2 family members (Bad, Bak, Bax, Bid, Bik, Bim_EL, Bok, Noxa, A1, Bcl-2, Bcl-w, Bcl-x_L and Mcl-1). Abnormalities in the apoptosis intrinsic pathway might impair clonal deletion of autoreactive T cells during thymic selection [13],[15]–[17]. In the periphery, autoreactive T cells can be silenced by anergy or deleted by apoptosis through activation-induced cell death (AICD), in which proteins from the death receptors family take part (extrinsic pathway: Fas, FasL and cFLIP) [18].

The process of pancreatic β cell damage involves a large number of autoreactive lymphocytes, and the mechanisms leading to the escape of those cells from tolerance mechanisms remain unclear [13],[19]. Some studies have demonstrated the association between apoptosis deregulation and defective thymic selection in AIDs [11],[13].

Several animal models of diseases which present impairment in the apoptosis process are associated with AIDs, such as non-obese diabetic (NOD) mice [20],[21]. NOD mice studies provide evidence that apoptosis deregulation represents an important mechanism of autoreactive T cell perpetuation in diabetes [21],[22]. Alteration in thymic negative selection and the AICD process has been described in NOD mice [21]–[23].

T1D mortality and morbidity emphasize the importance of therapeutic strategies to prevent and stabilize this autoimmune disorder. Traditional treatments include several daily exogenous insulin injections and glucose control, but although ameliorating the disease they do not stimulate the increase of the hormone physiological levels [24].

Many immunosuppressive interventions aimed at blocking immune reactions against pancreatic β cells have also been evaluated in patients with T1D [25]. Since 1996, severe AIDs have been treated successfully with high-dose immunosuppression followed by autologous non-myeloablative haematopoietic stem cell transplantation (HDI-AHSCT) [26]. Voltarelli and colleagues treated newly diagnosed patients T1D with this therapy in a clinical trial, with acceptable toxicity and with excellent results associated with the recovery endogenous insulin production [27]–[29]. However, some patients relapsed after a period of insulin independency, and the mechanisms involved in patients' response to HDI-AHSCT are not understood fully.

Besides the therapeutic efficacy of HDI-AHSCT in T1D patients, our group has been studying its immune mechanisms of action. Understanding of the remission/relapse immune mechanisms is fundamental for the establishment of HDI-AHSCT as a therapeutic option for T1D. This paper focuses on the determination of Bcl-2 gene expression, death receptor families and inhibitory apoptosis proteins (IAP) in T1D patients submitted to HDI-AHSCT. We discuss alterations in the intrinsic and extrinsic apoptosis pathways that might play a role in T1D pathogenesis. In addition, we associated the gene expression results with patients' clinical responses and correlated them with clinical laboratory parameters.

Materials and methods

Patients and controls

A total of 14 patients with newly diagnosed T1D, confirmed by glutamic acid decarboxylase autoantibody (GAD65) serum levels, were treated with HDI-AHSCT at the bone marrow transplantation unit of the School of Medicine of Ribeirão Preto, Brazil. Peripheral blood of T1D patients was obtained at five time-points during the study: pre-transplantation (pre-Tx), 180 days (D+180), 360 days (D+360), 540 days (D+540) and 720 days (D+720) after HDI-AHSCT. A detailed clinical trial protocol of HDI-AHSCT therapy for newly diagnosed T1D patients has been described (trial registration in clinicaltrials.gov identifier: NCT00315133) [27]–[29]. The control group for the gene expression studies comprised 14 healthy subjects paired with patients according to gender and age.

Table 1 summarizes patients' demographic features and clinical laboratory parameters at onset and D+720 post-HDI-AHSCT periods. This study was approved by the Ethics Committee from the University Hospital of the School of Medicine of Ribeirão Preto (process number 14105/06), and informed consent was obtained from T1D patients and controls before blood collection.

Table 1.

Patients' demographic features, clinical laboratory parameters at onset [pre-high-dose immunosuppression followed by autologous haematopoietic stem cell transplantation (HDI-AHSCT)] and at D+720 post-HDI/AHSCT and controls' gender and age.

Controls gender/age (years)	Patients	Gender/age (years)	Glycosylated haemoglobin (%) at onset/D+720	GAD65 autoantibodies (UI/ml) at onset/D+720	C-peptide peak stimulated (ηg/ml) at onset/D+720	Clinical evaluation at D+720 post-HDI/AHSCT
M/32	01	M/28	7·5/6·5	49·0/17·0	0·19/0·35	Insulin-free
M/23	02	M/21	9·3/5·3	1·1/0·0	0·32/0·59	Insulin-free
M/14	03	M/15	8·0/5·7	22·0/17·0	0·94/2·29	Insulin-free
M/17	04	M/16	7·7/5·5	51·0/46·0	1·04/2·81	Insulin-free
M/18	05	M/16	7·3/5·5	17·0/2·9	0·60/2·64	Insulin-free
F/21	06	F/20	10·0/6·9	4·0/17·0	0·03/1·46	Resumed-insulin*
M/19	07	M/17	5·4/6·8	48·0/5·6	0·15/1·80	Insulin-free
F/20	08	F/18	6·7/5·2	102·0/70·0	0·13/3·95	Insulin-free
F/19	09	F/17	8·9/6·2	44·0/13·0	0·43/n.d.	Resumed-insulin*
M/18	10	M/16	5·4/11·4	11·0/ND	0·39/n.d.	Resumed-insulin*
F/18	11	F/15	8·1/5·8	11·0/2·1	0·11/4·70	Insulin-free
M/24	12	M/24	8·1/5·7	24·0/0·0	1·03/1·65	Insulin-free
M/31	13	M/31	7·8/6·1	37·0/33·0	0·53/5·11	Insulin-free
M/19	14	M/17	10·1/5·9	21·1/7·6	0·28/1·83	Resumed-insulin*

^*Patients were transiently insulin-free, therefore they were classified as resumed-insulin group. Reference values: Hb A_1C < 6·5%; positive GAD65 > 1·0 U/ml; C-peptide 0·5–3·0 ηg/ml. F: female; GAD65, glutamic acid decarboxylase; Hb A_1C: glycosylated haemoglobin; M: male; n.d.: not determined.

Two years after HDI-AHSCT therapy, 10 patients were still insulin-free and four had resumed low doses of insulin. Furthermore, a significant increase in C-peptide levels was detected, indicating that the procedure was effective to enhance pancreatic insulin production [27]–[29]. The majority of adverse effects related to treatment were mild and included nausea, vomiting, fever and alopecia. With regard to severe adverse effects, two patients presented bilateral nosocomial pneumonia that responded completely to intravenous broad-spectrum antibiotics. During long-term follow-up, three patients presented with Graves' disease that was remitted completely with tapazol. One patient presented with rhabdomyolysis secondary to autoimmune hypothyroidism and one patient presented with both transient hypergonadotrophic hypogonadism and mild systemic lupus erythematosus. There was no mortality [27]–[29].

Peripheral blood mononuclear cell isolation

Peripheral blood mononuclear cells (PBMCs) from patients and controls were isolated from heparinized blood (20 ml) by centrifugation on a Ficoll-hypaque density gradient (d = 1·077; Amersham-Pharmacia, Uppsala, Sweden) [30]; 1·0 × 10⁷ cells were resuspended in a guanidine isothyocianate phenol solution (TRIzol®; Gibco BRL Life Technologies, Grand Island, NY, USA) and frozen at −80°C until RNA extraction.

RNA extraction and cDNA synthesis

Total RNA was isolated according to TRIzol® methodology (Gibco BRL Life Technologies) [31] and quantified spectrophotometrically. The concentration of RNA was adjusted to 1·0 µg/µl and cDNA was synthesized using 2·0 µg of RNA by employing a high-capacity cDNA archive kit (Applied Biosystems, Foster City, CA, USA) following the manufacturer's protocol. cDNA synthesis reaction was performed at 25°C for 10 min and then at 37°C for 2 h on a 9700 GeneAmp polymerase chain reaction (PCR) system (Applied Biosystems, Norwalk, CT, USA). The quality and integrity of the RNA samples are shown in Supporting information Table S1 and Supporting information Fig. S1.

Relative quantification of gene expression by real-time PCR

Relative quantification of anti-apoptotic genes a1, bcl-2, bcl-w, bcl-x_L and mcl-1 (Bcl-2 family); cIAP-1 and cIAP-2 (IAP family); extrinsic pathway gene c-FLIP_L and pro-apoptotic genes bad, bak, bax, bid, bik, bim_EL, bok and noxa (Bcl-2 family), fas and fasL (death receptor family) was performed by SYBR® Green PCR Master Mix Kit (Applied Biosystems, Foster City) on a 7500 real-time PCR system (Applied Biosystems, Foster City).

The PCR mixture consisted of 4·0 ng of cDNA, 10·0 µM of forward and reverse primers, 7·5 µL of SYBR® Green PCR Master Mix and 4·5 µl of deionized water to a final volume of 15 µl. The PCR conditions were: one cycle at 50°C for 2 min, 95°C for 10 min and 50 cycles at 95°C for 15 s, 54–62°C for 25 s (annealing temperatures were determined for each gene) and 72°C for 34 s.

For detection of anti-apoptotic and pro-apoptotic gene expression, we used the sequence primers described in Table 2. The β-actin and gapdh genes were used as housekeeping genes and the relative expression of the studied target genes were obtained after normalizing using the geometric average of the housekeeping gene mRNA levels. All reactions were duplicated and gene expression was calculated using the relative expression units (REU) method [32].

Table 2.

Primer sequences, amplicon size, and annealing temperature of apoptosis-related genes.

mRNA targets	Oligonucleotides (5′–3′)	Amplicon (bp)	AT (°C)
bad	F: CCG AGT GAG CAG GAA GAC TC	209 bp	60
	R: GGT AGG AGC TGT GGC GAC T
bak	F: TTT TCC GCA GCT ACG TTT TT	201 bp	60
	R: TGG TGG CAA TCT TGG TGA AGT
bax	F: CCC TTT TGC TTC AGG GTT TC	500 bp	56
	R: TCT TCT TCC AGA TGG TGA GTG
bid	F: GCT TCC AGT GTA GAC GGA GC	203 bp	62
	R: GTG CAG ATT CAT GTG TGG ATG
bik	F: TCT GCA ATT GTC ACC GGT TA	193 bp	60
	R: TTG AGC ACA CCT GCT CCT C
bimEL	F: TCT GCA ATT GTC ACC GGT TA	196 bp	59
	R: AAG ATG AAA AGC GGG GCT CT
bok	F: TTT TCC GCA GCT ACG TTT TT	249 bp	59
	R: TGG TGG CAA TCT TGG TGA AGT
noxa	F: AGC TGG AAG TCG AGT GTG CT	167 bp	54
	R: ACG TGC ACC TCC TGA GAA AA
a1	F: GGC TGG CTC AGG ACT ATC	227 bp	50
	R: CCA GTT AAT GAT GCC GTC
bcl-2	F: ACG AGT GGG ATG CGG GAG ATG TG	241 bp	67
	R: GCG GTA GCG GCG GGA GAA GTC
bcl-w	F: AGT TCG AGA CCC GCT TCC	308 bp	60
	R: CCC GTC CCC GTA TAG AGC
bcl-xL	F: CTG AAT CGG AGA TGG AGA CC	211 bp	60
	R: TGG GAT GTC AGG TCA CTG AA
mcl-1	F: AGA AAG CTG CAT CGA ACC AT	183 bp	56
	R: CC AGC TCC TAC TCC AGC AAC
fas	F: CAA GGG ATT GGA ATT GAG GA	203 bp	55
	R: TGG AAG AAA AAT GGG CTT TG
fasL	F: AGG AAA GTG GCC CAT TTA AC	176 bp	55
	R: CAA GAT TGA CCC CGG AAG TA
cFLIPL	F: GCC GAG GCA AGA TAA GCA	464 bp	54
	R: GCC CAG GGA AGT GAA GGT
cIAP-1	F: AGT CTT GCT CGT GCT GGT TT	550 bp	59
	R: ATG GAC AGT TGG GAA AAT GC
cIAP-2	F: AGT CTT GCT CGT GCT GGT TT	432 bp	58
	R: TGC TTT TGC CAG ATC TGT TG
β-actin	F: GCC CTG AGG CAC TCT TCC A	192 bp	58
	R: CCA GGG CAG TGA TCT CCT TCT
GAPDH	F: GCC TCA AGA TCA TCA GCA ATG C	111 bp	60
	R: CAT GGA CTG TGG TCA TGA GTC CT

AT: annealing temperature; bp: base pairs; F: forward primer; R: reverse primer; bp: base pairs.

Immunophenotypical analysis

PBMCs from patients and controls (5 ml) were collected in ethylenediamine tetraacetic acid (EDTA) (8·55 mg/tube) to perform blood counts and PBMC isolation. One hundred millilitres of peripheral mononuclear cell suspension (1·0 × 10⁶) were labelled with 5 µl of appropriated monoclonal antibodies anti-CD3, anti-CD4, anti-CD8, anti-FAS and anti-FASL (BD Pharmingen, San Jose, CA, USA) and incubated for 15 min at 4°C in the dark. After incubation, the cells were centrifuged for 5 min at 500 g, washed twice with fluorescence activated cell sorter (FACS) buffer (phosphate-buffered saline, 0·2% fetal bovine serum, 0·02% sodium azide), resuspended in 200 µl of FACS buffer and analysed by flow cytometry. Fifty thousand mononuclear cells were acquired using a FACS Canto Flow Cytometer (Becton-Dickinson, San Diego, CA, USA) and analysed by dot-plot using Diva version 6·0 software (Becton-Dickinson). Immunophenotyping of the CD3⁺CD4⁺Fas⁺, CD3⁺CD8⁺Fas⁺, CD3⁺CD4⁺FasL⁺ and CD3⁺CD8⁺FasL⁺ T subsets was performed and the results are shown as absolute cell numbers (cells/µl).

Protein extraction and Western blotting analysis

Total proteins were isolated according to TRIzol® sequential extraction methodology (Gibco BRL Life Technologies) [31]. Protein samples were diluted in sodium dodecyl sulphate 1%, quantified by the bicinchoninic acid (BCA) method using the BCA Protein Assay Kit (Pierce, Rockford, IL, USA) [33] and stored at −80°C. A total of 30 µg of protein was separated using 10 or 15% polyacrylamide gel electrophoresis and blotted onto polyvinylidene difluoride (PVDF) membranes using standard procedures. Following overnight incubation at 4°C in a blocking cocktail (5% light milk powder, Tris-HCl pH 7·5, 0·1% Tween 20 buffer), membranes were incubated with specific monoclonal primary antibodies for 24 h at 4°C.

We used monoclonal antibodies anti-Bcl-2 (BD Pharmingen), polyclonal rabbit antibodies anti-Bak (Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-Bim (BD Pharmingen), anti-Bcl-x (BD Pharmingen) and anti-FLICE-inhibitory protein (FLIP) (Calbiochem, La Jolla, CA, USA). Primary antibodies were diluted at 1:1000 in a blocking buffer plus 0·01% sodium azide. Following three washes in Tris-buffered saline (TBS)–Tween buffer, the membranes were labelled for 45 min with secondary antibody conjugated with horseradish peroxidase anti-mouse or anti-rabbit (GE Healthcare, Little Chalfont, Bucks, UK). Specific bands were detected using the enhanced chemiluminescence (ECL)-plus system (GE Healthcare) and quantified densitometrically using Alphaease software (Alpha Innotech, San Leandro, CA, USA).

Proteins such as polyclonal rabbit antibody anti-γ-tubulin and monoclonal anti-actin were used as endogenous controls. Results are presented by ratio of integrated density value (IDV) of specific proteins and from internal IDV control.

Statistical analysis

The Mann–Whitney t-test was used to compare gene expression between T1D patients' PBMCs and controls, controls and patients at D+720 post-AHSCT, and insulin-free and resumed-insulin patients. The differences in gene expression among T1D patients' PBMCs at pre-transplantation and several time-points post-HDI-AHSCT were made by Friedman test followed by Dunns' post-test. Spearman's test was used to detect correlations between gene expression results and patients' clinical laboratory parameters, such as anti-GAD65 and C-peptide levels. Tests with a P-value lower than 0·05 were considered statistically significant. Results were analysed using GraphPad Prism, version 5·0.

Results

Apoptosis-related genes expression profile in T1D patients

To analyse the pro-apoptotic and anti-apoptotic gene expression profile in T1D patients, we evaluated the expression of the genes from intrinsic and extrinsic apoptotic pathways and from IAP family genes in PBMCs isolated from patients and controls. We observed a significant down-regulation (P < 0·05) in pro-apoptotic bad (median: 0·066), bax (0·298) and fasL (6·101) expression in patients' PBMCs when compared to controls (median bad: 1401; bax: 4·194; fasL: 21·55) (Figs 1a–c and 8a) and an up-regulation in bid (0·552), bok (2·543) and noxa (9·516) gene expression in T1D patients in relation to controls (median bid: 0·008; bok: 0·516; noxa: 3·894) (Figs 1d–f and 8a). Conversely, we detected a significant increase (P < 0·05) in anti-apoptotic genes a1 (108·0), bcl-x_L (252·9) and cIAP-2 (257·7) in patients' PBMCs compared to controls (median a1: 66·91; bcl-x_L: 136·3; cIAP-2: 66·99) (Figs 2a–c and 8a) and a decrease in cIAP-1 (7·778) gene expression compared to controls (median cIAP-1: 64·19) (Figs 2d and 8a). There were no differences in bak, bik, bim_EL fas, bcl-2, bcl-w and cFLIP gene expression between T1D patients and controls (data not shown). Figure 8a summarizes the results obtained when gene expression data were compared between patients and controls.

Fig. 1.

Apoptosis-related pro-apoptotic gene expression profile in type 1 diabetes (T1D) patients; (a–c) bad, bax and fasL expression was down-regulated in T1D patients' peripheral blood mononuclear cells (PBMCs) (n = 14) in comparison to controls (n = 14); (d–f) bid, bok and noxa expression was up-regulated in T1D patients' PBMCs (n = 14) in comparison to controls (n = 14). Statistical analysis was performed by Mann–Whitney t-test. The box-plots show the median (horizontal bars), standard deviation, lower and upper quartiles. P-values lower than 0·05 were considered statistically significant.

Fig. 2.

Apoptosis-related anti-apoptotic gene expression profile in type 1 diabetes (T1D) patients; (a–c) a1, bcl-x_L and cIAP-2 expression was up-regulated in T1D patients' peripheral blood mononuclear cells (PBMCs) (n = 14) in comparison to controls (n = 14); (d) cIAP-1 expression was down-regulated in T1D patients' PBMCs (n = 14) in comparison to controls (n = 14). Statistical analysis was performed by Mann–Whitney t-test. The box-plots show the median (horizontal bars), standard deviation, lower and upper quartiles. P-values lower than 0·05 were considered statistically significant.

Modulation of apoptosis-related gene expression in T1D patients by HDI-AHSCT therapy

Once we observed a deregulation in several apoptotic genes in T1D patients' PBMCs, we analysed the effect of HDI-AHSCT therapy in the expression of apoptotic genes in T1D patients.

We noticed an increase in pro-apoptotic fas (median pre-Tx: 67·84; P < 0·05) and fasL (median pre-Tx: 6·101; P < 0·0001) gene expression in post-HDI-AHSCT periods, mainly at D+180 (median fasL: 33·79) and D+360 (fas: 174·3; fasL: 20·91) (Figs 3a–b and 8b). In contrast, a significant decrease in anti-apoptotic bcl-x_L (median pre-Tx:252·9; P = 0·001) gene expression at D+540 (median: 56·96) and D+720 (100·0) post-HDI-AHSCT periods was detected (Figs 3c and 8b).

Fig. 3.

Modulation of apoptosis-related gene expression in type 1 diabetes (T1D) patients by high-dose immunosuppression followed by autologous haematopoietic stem cell transplantation (HDI-AHSCT) therapy; (a,b) fas and fasL expression was up-regulated in T1D patients' peripheral blood mononuclear cells (PBMCs) at D+360 post-HDI/AHSCT (n = 14) in comparison to pre-HDI-AHSCT (n = 14); (c) bcl-x_L expression was down-regulated in T1D patients' PBMCs at D+540 post-HDI/AHSCT (n = 14) in comparison to pre-HDI-AHSCT (n = 14). Statistical analysis was performed by Friedman followed by Dunns' post-test. The box-plots show the median (horizontal bars), standard deviation, lower and upper quartiles. P-values lower than 0·05 were considered statistically significant.

Furthermore, to analyse whether pro-apoptotic and anti-apoptotic gene expression presented mRNA levels similar to controls after HDI-AHSCT therapy, we compared gene expression between T1D patients at D+720 post-HDI-AHSCT and controls. We found levels of bad, bax, bok and fasL pro-apoptotic gene expression similar to controls (Fig. 8b). The expression of bid, fas and noxa genes was modulated during follow-up; however, this expression was not similar to controls at D+720 post- HDI-AHSCT (data not shown).

In relation to anti-apoptotic gene expression, we observed a re-establishment of bcl-x_L and cIAP-1 gene expression levels, similar to that found in controls, in patients' PBMCs at 720 post-HDI-AHSCT (Fig. 8b). After HDI-AHSCT therapy, the expression of a1 and cIAP-2 genes were not similar to controls (data not shown). Figure 8b summarizes the analysis of gene expression results in T1D patients before and after the HDI-AHSCT therapy.

Expression of noxa was increased in insulin-free patients after HDI-AHSCT therapy

To analyse whether the pro-apoptotic and anti-apoptotic gene expression of T1D patients' PBMCs is associated with clinical response, we separated patients into two groups: insulin-free and resumed-insulin. We observed a significant increase (P < 0·05) in pro-apoptotic noxa gene expression in insulin-free patients (median: 18·95) at D+540 post-HDI-AHSCT when compared to patients who resumed insulin (median: 8·627) (Fig. 4).

Fig. 4.

Expression of noxa was increased in insulin-free patients after high-dose immunosuppression followed by autologous haematopoietic stem cell transplantation (HDI-AHSCT) therapy. Gene expression of noxa in insulin-free and resumed-insulin type 1 diabetes (T1D) patients' peripheral blood mononuclear cells (PBMCs) at D+540 post-HDI/AHSCT. Samples from 10 insulin-free and four resumed-insulin patients were analysed. Statistical analysis was performed by Mann–Whitney t-test. The box-plots show the median (horizontal bars), standard deviation, lower and upper quartiles. P-values lower than 0·05 were considered statistically significant.

Apoptosis-related protein expression in T1D patients

To evaluate whether pro-apoptotic and anti-apoptotic gene expression analysis correlated with protein expression, we selected genes based on the gene expression analysis and tested the proteins isolated from controls' and patients' PBMCs by Western blotting and flow cytometry.

We found no differences in Bak and Bim_EL protein expression when we compared them with samples isolated from T1D patients at pre-HDI-AHSCT versus controls and when we analysed pre-HDI-AHSCT versus D+360 post-HDI-AHSCT, similar to those obtained by gene expression (Fig. 5a,b).

Fig. 5.

Apoptosis-related protein expression in controls and in type 1 diabetes (T1D) patients. (a,b) Bak and Bim pro-apoptotic protein expression in controls (n = 5) and T1D patients at pre-high-dose immunosuppression followed by autologous haematopoietic stem cell transplantation (HDI-AHSCT) (n = 5) and at D+360 post-HDI-AHSCT (n = 5). (c–e) Bcl-2, Bcl-x_L and cellular FLICE-inhibitory protein (cFLIP)_L anti-apoptotic protein expression in controls and T1D patients at pre-transplant and D+360 post-HDI-AHSCT.

Regarding anti-apoptotic molecules, Bcl-2 protein expression in T1D patients' PBMCs at pre-HDI-AHSCT (IDV: 1·15 ± 1·12) was up-regulated when compared to protein content obtained from controls (0·68 ± 0·29) and increased in D+360 (1·85 ± 0·61) post-HDI-AHSCT (Fig. 5c). Bcl-x_L protein expression was augmented in patients' PBMCs during the pre-HDI-AHSCT period (1·43 ± 0·49) compared to the control group (1·24 ± 0·15) and decreased at D+360 post-HDI-AHSCT (0·72 ± 0·27) (Fig. 5d).

In accordance with the results obtained by real-time PCR, c-FLIP_L protein expression was similar in patients' PBMCs at pre-HDI-AHSCT (0·81 ± 0·48) and in controls (0·92 ± 0·06). However, there was a reduced protein expression at D+360 post-HDI-AHSCT therapy, unobserved in gene expression analysis (0·43 ± 0·08) (Fig. 5e).

Once we had observed an increased expression of fas and fasL pro-apoptotic genes in patients' PBMCs during the first year post-HDI-AHSCT therapy, we analysed the absolute numbers of CD3⁺CD4⁺ and CD3⁺CD8⁺ T subsets expressing Fas and FasL by flow cytometry. We found a significant decrease in the absolute numbers of CD3⁺CD4⁺Fas⁺ T cells at D+180 post-HDI-AHSCT compared to pre-HDI-AHSCT (Fig 6a). No differences in the absolute numbers of CD3⁺CD8⁺Fas⁺, CD3⁺CD4⁺FasL⁺ and CD3⁺CD8⁺FasL⁺ T subsets were observed when we compared samples from T1D patients at pre-HDI-AHSCT versus controls and between patients at pre-HDI-AHSCT versus post-HDI-AHSCT periods (Fig. 6b–d).

Fig. 6.

Absolute numbers of CD3⁺CD4⁺ and CD3⁺CD8⁺ T cells expressing Fas or FasL molecules. The absolute number of CD4⁺ and CD8⁺ T cells expressing Fas or FasL molecules was determined by flow cytometry. (a,b) Absolute numbers of CD3⁺CD4⁺Fas⁺ and CD3⁺CD8⁺Fas⁺ T lymphocytes in controls (n = 10) and type 1 diabetes (T1D) patients at pre-high-dose immunosuppression followed by autologous haematopoietic stem cell transplantation (HDI-AHSCT) (n = 9), D+180 (n = 7) and at D+360 post-HDI-AHSCT (n = 5). (c,d) Absolute numbers of CD3⁺CD4⁺FasL⁺ and CD3⁺CD8⁺FasL⁺ T lymphocytes in controls (n = 10) and T1D patients at pre-HDI-AHSCT (n = 9), D+180 (n = 7) and at D+360 post-HDI-AHSCT (n = 5). Statistical analysis was performed by Mann–Whitney t-test. The box-plots show the median (horizontal bars), standard deviation, lower and upper quartiles. P-values lower than 0·05 were considered statistically significant.

Correlation between patients' clinical laboratory parameters and gene expression results

Once we had observed a gene expression modulation in patients after the HDI-AHSCT therapy, we proposed to analyse the correlations between pro-apoptotic and anti-apoptotic gene expression and the T1D patients' clinical laboratory parameters (GAD65 and C-peptide levels).

Down-regulation of pro-apoptotic fasL expression in patients' PBMCs correlated positively with C-peptide levels at pre-HDI-AHSCT (P = 0·010; r = 0·60) (Fig. 7a). Conversely, pro-apoptotic noxa expression at D+360 post-HDI-AHSCT correlated with GAD65 levels (P = 0·037; r = −0·49), whereas anti-apoptotic cIAP-1 expression at D+720 post-HDI-AHSCT correlated with GAD65 levels (P = 0·041; r = 0·50) (Fig. 7b,c).

Fig. 7.

Patients' clinical laboratory parameters and gene expression correlation. (a) Decreased pro-apoptotic fasL gene expression correlated positively with C-peptide levels at pre-high-dose immunosuppression followed by autologous haematopoietic stem cell transplantation (HDI-AHSCT). (b) Increased pro-apoptotic noxa gene expression correlated conversely with glutamic acid decarboxylase autoantibodies (GAD65) levels at D+360 post-HDI/AHSCT. (c) Decreased anti-apoptotic cIAP-1 gene expression correlated positively with GAD65 levels at D+720 post-HDI/AHSCT. Samples from 14 type 1 diabetes mellitus (T1DM) patients, collected at different time-points, were analysed. Statistical analysis was performed by Spearman's correlation and P-value lower than 0·05 was considered statistically significant.

Discussion

Apoptosis plays an important role during the development of B and T lymphocytes during negative selection in bone marrow and thymus, respectively, and in clonal deletion during peripheral tolerance [15],[16]. Thus, impairment in the apoptosis process during the central and peripheral tolerance mechanisms might result in the escape of autoreactive lymphocytes and in autoimmune disease development [13],[14],[34].

Although the pathogenesis of autoimmune diseases involves several factors and was not understood completely, previous studies have demonstrated the involvement of abnormal apoptosis in autoimmune disease development, including T1D [11]–[14],[35]. Strong evidence that diabetic patients have altered expression of apoptosis genes during thymic selection and the mechanisms of AICD are derived from studies performed in NOD mice [13],[21]–[23]. Few studies have reported that the thymic selection and apoptotic processes are defective in diabetic patients after an immune response [12],[36]. The relevance of our present work lies in confirming the involvement of defective apoptosis in T1D pathogenesis in humans, demonstrating the modulation of apoptosis gene expression after HDI-AHSCT therapy, the association of patients' clinical response (insulin-free or resumed-insulin) and the clinical laboratory parameters to gene expression results.

The mechanism of apoptosis control involves the balance between pro-apoptotic and anti-apoptotic molecules, particularly the cellular concentrations of Bcl-2 or Bcl-X_L and Bax or Bad [34],[37],[38]. Thus, it is important to understand how the decrease in pro-apoptotic genes and increase in the anti-apoptotic gene expression from Bcl-2 family in T1D patients' PBMCs contribute to apoptotic resistance of pathogenic autoreactive lymphocytes. The increased expression of anti-apoptotic proteins Bcl-2 and Bcl-x_L in patients' PBMCs also suggests the contribution of altered apoptosis in T1D pathogenesis. These molecules may have a possible role in the perpetuation of autoreactive cells by abnormal clonal deletion during immune tolerance processes. Previous studies in NOD mice showed an up-regulated expression of the anti-apoptotic genes bcl-2 and bcl-x_L in thymocytes derived from these animals, contributing to the increased resistance of immature thymocytes to apoptosis and defective deletion of autoreactive thymocytes and their perpetuation in periphery [39]–[41].

Although there are few data regarding apoptotic gene expression linked to T1D pathogenesis in humans, altered expression of apoptosis-related genes has been described in other autoimmune diseases (AIDs), such as decreased expression of bad and bax and increased expression of bcl-2 and bcl-x_L in lymphocytes from multiple sclerosis patients [42]–[44] and rheumatoid arthritis [45],[46], and increased bcl-2 expression in PBMCs isolated from systemic lupus erythematosus patients [47],[48]. In those studies, the deregulated apoptosis process in immune cells has been implicated in the development of AID [42]–[48]. Furthermore, apoptosis studies both in human [49] and mouse with AIDs [23],[41] have demonstrated abnormal expression of the molecules from the extrinsic pathway involving fas and fasL in lymphocytes. They also suggest that the breakdown of immune tolerance to self-antigens could be associated with defective AICD during peripheral tolerance processes. The IAP proteins prevent apoptosis by direct inhibition of activation of pro-caspases 3, 7 and 9 in active caspases [50], and the increased gene expression of cIAP-2 in T1D patients' PBMCs also suggests a possible involvement of this molecule in the maintenance of autoreactive cells in the periphery and therefore a higher resistance to apoptosis.

We showed a large-scale deregulation of the apoptotic molecules that participate in several key points of the apoptotic pathways, as shown in Fig. 8a. In the extrinsic pathway, we observed reduced expression of pro-apoptotic fasL, indicating an impairment of the AICD process mediated by Fas/FasL. In the intrinsic apoptosis pathway, we showed increased expression of anti-apoptotic molecules (a1 and bcl-x_L), which send signals to the mitochondria to inhibit the activation of pro-apoptotic proteins involved in mitochondrial permeability, such as bad and bax, which showed decreased expression. These changes, as well as in other molecules of the Bcl-2 family, may affect the subsequent apoptotic processes (release of cytochrome c and activation of effector caspases). Taken together, our results suggest that deregulation of expression of apoptosis-related genes in T1D patients is involved in the maintenance of autoreactive lymphocytes and in T1D pathogenesis.

Fig. 8.

Schematic drawing of the extrinsic and intrinsic pathways of apoptosis. (a) Alterations in gene expression in peripheral blood mononuclear cells (PBMCs) isolated from type 1 diabetes (T1D) patients compared to control subjects. Red arrows indicate genes up-regulated in patients and green arrows indicate genes down-regulated in patients compared to control group. (b). Alterations in gene expression in patients' PBMCs after high-dose immunosuppression followed by autologous haematopoietic stem cell transplantation (HDI-AHSCT) therapy compared to pre-HDI-AHSCT. Red arrows indicate genes up-regulated in patients after the HDI-AHSCT and green arrows indicate genes down-regulated in patients after the HDI-AHSCT therapy. The parentheses indicate the period in which the genes were altered. Blue boxes indicate anti-apoptotic molecules and green boxes indicate pro-apoptotic molecules. Stars indicate the genes modulated during follow-up, and had expression levels similar to controls at D+720 post-HDI-AHSCT.

The rationale of the HDI-AHSCT therapy for AIDs is that high-dose immunosuppression (cyclophosphamide and rabbit anti-thymocyte globulin) eliminates most of the T and B cell repertoire, including the autoreactive pathogenic lymphocytes, and promotes reconstitution of a new immune system from haematopoietic stem cell precursors which could restore the tolerance process (immune resetting), thereby preventing disease progression [51]–[53]. Some studies have shown the role of immunosuppression with polyclonal rabbit anti-thymocyte globulin in the elimination of autoreactive lymphocytes and blocking the autoimmune attack on pancreatic beta cells by promoting an increase in C-peptide levels and a decrease in daily insulin doses in newly diagnosed T1D patients [54]–[56]. Our data indicate that HDI-AHSCT therapy was able to modulate the apoptotic gene expression in mononuclear cells from T1D patients towards levels similar to healthy controls. We suggest that re-establishment of the normal apoptotic gene expression profile after transplantation, together with the increased expression of pro-apoptotic genes (fas and fasL) and decreased expression of anti-apoptotic bcl-x_L (as summarized in Fig. 8b), might constitute one of the mechanisms involved in the re-establishment of immune tolerance in T1D patients who responded to HDI-AHSCT therapy.

The AICD represents a major mechanism of peripheral tolerance induction and plays an important role in the deletion of autoreactive lymphocytes that escape from central tolerance [23],[36]. The increase in expression of the pro-apoptotic genes fas and fasL in PBMCs isolated from T1D patients after HDI-AHSCT suggest a re-establishment of tolerance mechanisms mediated by AICD and an increase in the apoptosis susceptibility of residual autoreactive lymphocytes. These results are in agreement with previous observations in the literature that showed up-regulation of the Fas molecule on immune cells after immunosuppression with anti-thymocyte globulin (ATG) treatment [57]–[59] and an increased frequency of CD4⁺Fas⁺ and CD8⁺Fas⁺ cells in multiple sclerosis patients during the first months of immune reconstitution after AHSCT [60].

The remission of T1D patients after HDI-AHSCT correlates with decreased GAD65 levels and increased C-peptide levels [27],[29]. The correlation of decreased GAD65 levels with increased gene expression of pro-apoptotic noxa suggests a possible role of this pro-apoptotic gene in the positive therapeutic response of T1D patients to HDI-AHSCT. Conversely, at the pre-transplantation period, the observation that decreased C-peptide levels correlate positively with decreased fasL expression possibly indicates the importance of this apoptotic molecule in peripheral tolerance breakdown and autoreactive lymphocyte maintenance in T1D patients.

Taken together, our data suggest a key role of deregulated apoptosis in the development of T1D and in the perpetuation of autoreactive lymphocytes in patients. We suggest that abnormal apoptosis, mainly the decrease in pro-apoptotic genes and the increase in anti-apoptotic gene expression, might be involved in central and peripheral tolerance breakdown and consequently in T1D pathogenesis. Furthermore, HDI-AHSCT therapy modulated apoptosis gene expression towards levels similar to controls, mainly at 2 years after transplantation. Moreover, it is noteworthy that re-establishment of the normal gene expression profile correlated with a positive response to HDI-AHSCT (insulin-free patients). Further studies are needed in order to use these apoptosis-related molecules as possible biomarkers of T1D clinical remission as well as targets for therapeutic interventions.

Disclosure

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Supporting information

Fig. S1. The electrophoretic profile of RNAisolated from peripheral blood mononuclear cells (PBMCs) of type Idiabetes (T1D) patients and controls. Lines 1–3 [RNA samplesof T1D patients at pre-high-dose immunosuppression followed byautologous haematopoietic stem cell transplantation (HDI-AHSCT)],lines 4–7 (RNA samples of T1D patients at post- HDI-AHSCTperiods) and lines 8–9 (RNA samples of controls).

Table S1. Quantification and quality of the RNAs samples from type I diabetes (T1D) patients (pre- and post HDI-AHSCT periods) and controls.