Two major unanswered questions related to the pathogenesis of type 1 diabetes (T1D) prevent progress in our ability to intervene in disease development: What breaks immune tolerance to the pancreatic β-cells? And what kills the β-cells?

By the surprising demonstration that knockout of the tryptase granzyme A (GrzmA), conventionally considered to be a key proapoptotic mediator of immune killer cell cytotoxicity, does not protect but markedly accelerates and increases incidence of T1D in the nonobese diabetic (NOD) mouse, Mollah et al. (1) add to answering both of these questions. On one hand, they show that GrzmA does not contribute to β-cell killing. On the other hand, their striking finding that diabetes resistance in the proinsulin II transgenic mouse is broken by GrzmA knockout provides evidence of an important role of GrzmA for maintenance of peripheral tolerance.

By answering “No” to the question as to whether GrzmA is important in β-cell killing, this study adds to reports showing that granzyme B (GrzmB) (2), Fas (3,4), and perforin (5,6) are dispensable in β-cell killing in the NOD mouse. These observations challenge the view that β-cell destruction in T1D is caused by classic effector T-cell mechanisms and support the notion that these cells do ensure antigen-specific homing and reactivity but then interact with innate immune cells to orchestrate an intraislet inflammatory reaction leading to β-cell killing, the selectivity of which is in part determined by differentiation-dependent sensitivity of β-cells to cytotoxic inflammatory mediators (7).

The “Yes” to the question as to whether GrzmA contributes to maintain peripheral tolerance is perhaps even more important because it identifies new potential therapeutic targets. T-cell GrzmA expression and production is reduced in human T1D by yet unknown mechanisms (8). Understanding of these mechanisms and restoration of GrzmA levels may have therapeutic perspectives. Key in this respect would be clarification of the regulation and mechanism of action of GrzmA.

GrzmA, discovered by Tschopp and colleagues (9) in 1986, belongs to the granzyme family of serine proteases encompassing 10 members in the mouse and 5 members in the human (10,11). They have trypsin-, chymotrypsin-, and elastase-like proteolytic activities, but in contrast to these broad serine proteases, granzymes have higher substrate specificity conferred by a substrate binding pocket. Granzymes are expressed mainly in cytotoxic T cells and natural killer (NK) cells in response to interleukin (IL)-2 in synergy with IL-12 (12,13), with GrzmA being the most abundantly expressed but also having lower cytotoxic activity than GrzmB. GrzmA and GrzmB are also expressed in γδT cells, a subset of the intraepithelial lymphocytes abundant in the intestinal mucosa, and in thymocytes, suggesting roles in both central tolerance and gut immunity. GrzmB, but as yet not GrzmA, is also found to be expressed in plasmacytoid dendritic cells (pDCs) (14), the innate master producer of the antiviral type I interferon (IFN) IFNα. Apart from mediating T-cell and NK-cell granule-dependent cytotoxicity, granzymes regulate B lymphocyte proliferation, inhibit viral replication, and kill bacteria.

In granule-dependent cytotoxicity granzymes are delivered to early endosomes by clathrin- and dynamin-dependent endocytosis after entering the target cell via perforin, a membrane pore-forming molecule cosecreted with granzymes (Fig. 1). Granzymes enter the cytosol from the endosomal compartment and are then activated by the near-neutral pH of the cytosol, where they in turn trigger the caspase-dependent intrinsic death pathway (11). Until recently it was believed that GrzmA was a mere “backup” for GrzmB in perforin-mediated apoptosis and activated identical death pathways. It is now clear that not only do GrzmA- and GrzmB-deficient mice have differential susceptibility to different infections, but GrzmA also triggers noncaspase-dependent cell death by activating the DNA-degrading SET complex (Fig. 1) (10,11). This activity relates to a unique ability of GrzmA to homodimerize, thereby exposing an extended exosite that determines its substrate specificity.

Figure 1

GrzmA activates a caspase-independent death pathway but prevents accumulation of ssDNA. GrzmA (closed scissors) is delivered through perforin (*) to the early endosome, is liberated from the endosome, and is activated (open scissors) by the neutral pH in the cytosol. It then first enters the mitochondrial matrix and next cleaves a component of the electron transport complex I NDUFS3, thereby perturbing mitochondrial redox function, ATP generation, and maintenance of the mitochondrial membrane potential, as well as causing reactive oxygen species (ROS) formation. ROS drive nuclear translocation of the endoplasmic reticulum (ER)-associated oxidative stress response complex SET that contains six components: three nucleases, the base excision repair endonuclease Ape1, the endonuclease NM23-H1, and the 5′–3′ exonuclease TREX1; the chromatin modifying proteins SET and pp32; and HMGB2, a DNA binding protein that recognizes distorted DNA. SET normally serves to repair abasic sites caused by oxidative damage in DNA. GrzmA translocating to the nucleus cleaves histone 1 and cuts histone tails. Hereby the chromatin structure is opened up allowing access to nucleases. GrzmA then cleaves Ape1, HMGB2, and SET, which otherwise binds and inhibits NM23-H1. Activated NM23-H1 in turn nicks DNA (pliers), which is further degraded by TREX1 (pliers), thereby preventing liberation of ssDNA fragments into the cytosol. Mollah et al. (1) suggest that in the absence of GrzmA, ssDNA fragments accumulate, are liberated, and activate Toll-like receptors on pDCs, triggering type I IFN production, which breaks immune tolerance.

Figure 1

GrzmA activates a caspase-independent death pathway but prevents accumulation of ssDNA. GrzmA (closed scissors) is delivered through perforin (*) to the early endosome, is liberated from the endosome, and is activated (open scissors) by the neutral pH in the cytosol. It then first enters the mitochondrial matrix and next cleaves a component of the electron transport complex I NDUFS3, thereby perturbing mitochondrial redox function, ATP generation, and maintenance of the mitochondrial membrane potential, as well as causing reactive oxygen species (ROS) formation. ROS drive nuclear translocation of the endoplasmic reticulum (ER)-associated oxidative stress response complex SET that contains six components: three nucleases, the base excision repair endonuclease Ape1, the endonuclease NM23-H1, and the 5′–3′ exonuclease TREX1; the chromatin modifying proteins SET and pp32; and HMGB2, a DNA binding protein that recognizes distorted DNA. SET normally serves to repair abasic sites caused by oxidative damage in DNA. GrzmA translocating to the nucleus cleaves histone 1 and cuts histone tails. Hereby the chromatin structure is opened up allowing access to nucleases. GrzmA then cleaves Ape1, HMGB2, and SET, which otherwise binds and inhibits NM23-H1. Activated NM23-H1 in turn nicks DNA (pliers), which is further degraded by TREX1 (pliers), thereby preventing liberation of ssDNA fragments into the cytosol. Mollah et al. (1) suggest that in the absence of GrzmA, ssDNA fragments accumulate, are liberated, and activate Toll-like receptors on pDCs, triggering type I IFN production, which breaks immune tolerance.

Having shown that enhanced diabetes development in GrzmA−/− NOD mice depended on type I IFN signaling and that innate immune cell accumulation of ssDNA was associated with a type I IFN–related islet gene expression signature, Mollah et al. (1) make the case that GrzmA deficiency permits liberation of ssDNA fragments into the cytosol in pDCs due to lack of GrzmA-mediated activation of the SET complex nuclease activity and that the ssDNA fragments subsequently promote type I IFN production (likely via Toll-like receptors) that in turn breaks tolerance (Fig. 1). However, it remains to be shown that pDCs express GrzmA, and GrzmA deficiency did not increase IFNα in pDC culture medium or in circulation. ssDNA was increased in GrzmA-expressing NK cells correlating with increased NK numbers, but NK cells are IFNγ—not type I IFN—producers. Although ssDNA did not accumulate in GrzmA-expressing T cells, β-cell antigen–specific CD8+ T cells were increased in islets, lymph nodes, and spleen. Regulatory T-cell number and function were unaffected. Taken together, these observations suggest that GrzmA may directly regulate the autoreactive T-cell pool. Thus GrzmB has been shown to mediate activation-induced cell death in T cells (15), and pDC-derived GrzmB suppresses effector T-cell proliferation by degrading the T-cell receptor ζ-chain (16). It would be relevant to determine whether GrzmA has similar actions.

What is the human relevance of this study? T1D risk is increased in clinical trials of type I IFN (17), underpinning that aberrant production of type I IFN may break tolerance. Yet targeting type I IFN alone does not prevent diabetes development in the NOD mouse (18). There are no human studies of anti-IFNα in T1D. The authors raise the caveat that NOD mice carry endogenous nonpathogenic retroviruses that may be the source of accumulating cytosolic ssDNA. However, common human diseases are caused by DNA viruses (adenovirus, herpes simplex virus, poxviruses, and hepatitis B virus), prompting human studies. Also since IL-2 deficiency is a central defect in T1D pathogenesis (19) and since GrzmA expression is driven by IL-2, the observation that T-cell GrzmA is reduced in children with T1D (8) further emphasizes the translational implications of the study by Mollah et al. (1). Larger clinical trials of IL-2 in T1D, now shown in small studies to be safe and yielding increases in regulatory T cells (20), are under way.

See accompanying article, p. 3041.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

1.
Mollah
ZUA
,
Quah
HS
,
Graham
KL
, et al
.
Granzyme A deficiency breaks immune tolerance and promotes autoimmune diabetes through a type I interferon–dependent pathway
.
Diabetes
2017
;
66
:
3041
3050
[PubMed]
2.
Kobayashi
M
,
Kaneko-Koike
C
,
Abiru
N
, et al
.
Genetic deletion of granzyme B does not confer resistance to the development of spontaneous diabetes in non-obese diabetic mice
.
Clin Exp Immunol
2013
;
173
:
411
418
[PubMed]
3.
Allison
J
,
Strasser
A
.
Mechanisms of β cell death in diabetes: a minor role for CD95
.
Proc Natl Acad Sci U S A
1998
;
95
:
13818
13822
[PubMed]
4.
Savinov
AY
,
Tcherepanov
A
,
Green
EA
,
Flavell
RA
,
Chervonsky
AV
.
Contribution of Fas to diabetes development
.
Proc Natl Acad Sci U S A
2003
;
100
:
628
632
[PubMed]
5.
Trivedi
P
,
Graham
KL
,
Krishnamurthy
B
, et al
.
Perforin facilitates beta cell killing and regulates autoreactive CD8+ T-cell responses to antigen in mouse models of type 1 diabetes
.
Immunol Cell Biol
2016
;
94
:
334
341
[PubMed]
6.
Amrani
A
,
Verdaguer
J
,
Anderson
B
,
Utsugi
T
,
Bou
S
,
Santamaria
P
.
Perforin-independent β-cell destruction by diabetogenic CD8+ T lymphocytes in transgenic nonobese diabetic mice
.
J Clin Invest
1999
;
103
:
1201
1209
[PubMed]
7.
Hansen
JB
,
Tonnesen
MF
,
Madsen
AN
, et al
.
Divalent metal transporter 1 regulates iron-mediated ROS and pancreatic β cell fate in response to cytokines
.
Cell Metab
2012
;
16
:
449
461
[PubMed]
8.
Hamel
Y
,
Mauvais
FX
,
Pham
HP
, et al
.
A unique CD8+ T lymphocyte signature in pediatric type 1 diabetes
.
J Autoimmun
2016
;
73
:
54
63
[PubMed]
9.
Masson
D
,
Zamai
M
,
Tschopp
J
.
Identification of granzyme A isolated from cytotoxic T-lymphocyte-granules as one of the proteases encoded by CTL-specific genes
.
FEBS Lett
1986
;
208
:
84
88
[PubMed]
10.
Trapani JA. Granzymes: a family of lymphocyte granule serine proteases. Genome Biol 2001;2:reviews3014.1–reviews3014.7
11.
Lieberman
J
.
Granzyme A activates another way to die
.
Immunol Rev
2010
;
235
:
93
104
[PubMed]
12.
Janas
ML
,
Groves
P
,
Kienzle
N
,
Kelso
A
.
IL-2 regulates perforin and granzyme gene expression in CD8+ T cells independently of its effects on survival and proliferation
.
J Immunol
2005
;
175
:
8003
8010
[PubMed]
13.
DeBlaker-Hohe
DF
,
Yamauchi
A
,
Yu
CR
,
Horvath-Arcidiacono
JA
,
Bloom
ET
.
IL-12 synergizes with IL-2 to induce lymphokine-activated cytotoxicity and perforin and granzyme gene expression in fresh human NK cells
.
Cell Immunol
1995
;
165
:
33
43
[PubMed]
14.
Rissoan
MC
,
Duhen
T
,
Bridon
JM
, et al
.
Subtractive hybridization reveals the expression of immunoglobulin-like transcript 7, Eph-B1, granzyme B, and 3 novel transcripts in human plasmacytoid dendritic cells
.
Blood
2002
;
100
:
3295
3303
[PubMed]
15.
Sharma
V
,
Delgado
M
,
Ganea
D
.
Granzyme B, a new player in activation-induced cell death, is down-regulated by vasoactive intestinal peptide in Th2 but not Th1 effectors
.
J Immunol
2006
;
176
:
97
110
[PubMed]
16.
Jahrsdörfer
B
,
Beyer
T
,
Schrezenmeier
H
,
Debatin
K-M
,
Fabricius
D
.
Granzyme B is a key regulator of plasmacytoid dendritic cell immunogenicity
.
Blood
2014
;
124
:
4127
17.
Schreuder
TC
,
Gelderblom
HC
,
Weegink
CJ
, et al
.
High incidence of type 1 diabetes mellitus during or shortly after treatment with pegylated interferon alpha for chronic hepatitis C virus infection
.
Liver Int
2008
;
28
:
39
46
[PubMed]
18.
Quah
HS
,
Miranda-Hernandez
S
,
Khoo
A
, et al
.
Deficiency in type I interferon signaling prevents the early interferon-induced gene signature in pancreatic islets but not type 1 diabetes in NOD mice
.
Diabetes
2014
;
63
:
1032
1040
[PubMed]
19.
Hulme
MA
,
Wasserfall
CH
,
Atkinson
MA
,
Brusko
TM
.
Central role for interleukin-2 in type 1 diabetes
.
Diabetes
2012
;
61
:
14
22
[PubMed]
20.
Hartemann
A
,
Bensimon
G
,
Payan
CA
, et al
.
Low-dose interleukin 2 in patients with type 1 diabetes: a phase 1/2 randomised, double-blind, placebo-controlled trial
.
Lancet Diabetes Endocrinol
2013
;
1
:
295
305
[PubMed]
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.