A New Tool for Dissecting Genetic Control of Type 1 Diabetes

  1. William M. Ridgway
  1. Division of Immunology, Allergy and Rheumatology, University of Cincinnati College of Medicine, Cincinnati, OH
  1. Corresponding author: William M. Ridgway, wridg1{at}

The NOD mouse has been a critical tool in the quest to understand the genetic control of type 1 diabetes (T1D), and over 25 murine insulin-dependent diabetes (Idd) loci that modulate the natural history of T1D have been identified (1). Several of the candidate genes identified in NOD mice also play a role in human T1D, suggesting that dysregulated immune pathways in NOD may closely resemble those found in humans, and justifying continuing work on the genetic origin of T1D in NOD (2). The role of specific genes in T1D has been explored by constructing congenic mice (carrying disease-protective Idd loci), by creating transgenic mice, and by knocking out genes. In each case, however, these studies have had a significant potential drawback: the presence of “passenger DNA” that can potentially confound the interpretation of the results (Fig. 1A, top panel) (3). Knockout mice, for example, have previously been made almost exclusively using non-NOD embryonic stem (ES) cells; the resulting non-NOD mice were backcrossed to NOD mice. The backcrossing process ensures that non-NOD genetic material is bred along with the genetic region of interest (Fig. 1A, top panel). This is not merely a theoretical concern, as some published studies of transgenic or deleted genes showed an effect on T1D that was later proven to arise from non-NOD passenger DNA (3).

Figure 1

A: Zinc-finger nuclease (ZFN) knockout technology eliminates passenger DNA. The top panel illustrates chromosomal DNA from a conventional knockout bred onto the NOD genetic background. Passenger DNA from black 6 (B6) and 129/SV (129) mouse strains is found in the NOD knockout mouse. White shading, NOD genetic material; black shading, B6 genetic material; hatched shading, 129 genetic material. The knockout itself is represented by the thin black line. The bottom panel shows a ZFN knockout on the NOD background: passenger DNA is eliminated. B: Biological complexity of CD137 signaling. A CD137 knockout not only affects signaling of CD137+ cells initiated by CD137L-expressing cells, but also eliminates CD137 signaling of CD137L-expressing cells (reverse signaling) initiated by CD137+ cells and the potential regulatory effect of soluble CD137 on immune cell activation. mCD137, membrane-bound CD137; sCD137, soluble CD137; APC, antigen-presenting cell. (A high-quality color representation of this figure is available in the online issue.)

In the current issue, Chen et al. (4) have eliminated this passenger DNA problem using NOD ES cells and zinc-finger nuclease technology to knock out a candidate gene, Cd137, in NOD mice without introducing non-NOD genetic material. A similar approach was used recently to understand the role of the class II–associated molecule DM, and taking advantage of an existing Balb/c construct in a region of genetic homology to NOD to eliminate DM in NOD ES cells (5). These studies represent breakthroughs for understanding specific genetic effects in NOD mice. The approach used by Chen et al. (4) takes advantage of zinc-finger nucleases targeted to specific genetic regions (in this case, exon 4 of Cd137); the nuclease cleaves the genome at the site of interest, and subsequent genome repair mechanisms excise the damaged region thus knocking out the gene (6). This resulted in a specific CD137 knockout on a 100% NOD genetic background (Fig. 1A, bottom panel).

Cd137 is an interesting candidate gene in the Idd9.3 region, and CD137 has profound effects on the immune system at many levels (7). In T1D, Lyons et al. (8) showed that the NOD Cd137 allele differed from B10 by three exonal single nucleotide polymorphisms. Cannons et al. (9) clearly demonstrated that the NOD CD137 allele was hypofunctional: CD137–mediated stimulation produced significantly less interleukin-2 and proliferation from NOD than NOD.B10 Idd9.3 T cells (expressing the protective B10 CD137 variant). This result, however, was somewhat confusing: how could hypofunctional NOD CD137 contribute to T1D? CD137 has multiple roles throughout the immune system (Fig. 1B); it is critical to T-cell effector function and to the acquisition of CD8 T-cell memory (7). T-regulatory cells (Tregs) also express CD137, and CD137-expressing Tregs were found in the pancreatic islet (1012). Thus CD137 in T1D might function in both effector T cell (CD8 cells mediating islet cell damage) or in regulation against autoimmunity (on Tregs). Indeed, Irie et al. (11) found that stimulation via CD137 could either diminish or augment T1D depending on the timing of treatment. Kachapati et al. (13) showed that the hypofunctional NOD allele was associated with significantly decreased numbers of CD137+ Tregs in NOD mice; that these Tregs produced soluble CD137, which was immunosuppressive; and that CD137+ Tregs are functionally superior to other Treg subsets. Thus, a hypofunctional allele could decrease protective immunity, allowing T-cell effector immunity to mediate tissue damage.

Interesting results found by Chen et al. (4) in their NOD CD137−/− mice raise additional questions about the role of CD137 in T1D and suggest that further work must be done before we understand this complicated system. They found that NOD CD137−/− mice had delayed onset of diabetes and decreased total diabetes incidence in the absence of significant differences in insulitis. Interestingly, CD4 and CD8 cells from NOD CD137−/− mice were markedly deficient in transferring T1D, even when CD25 Tregs were depleted from the transfer—suggesting that the CD137 knockout decreased the pathogenicity of T-cell effectors in this model. Nonetheless, there was also a suggestion that Tregs in CD137−/− mice were defective, as they were less suppressive in vitro.

How should we understand these results in the context of previous studies? Chen et al. (4) show that in the CD137−/−mice the effect on T-cell effectors may be the dominant effect: the loss of CD137 might be preventing the accumulation of diabetogenic CD8 effector/memory T cells. Still, it is surprising, given that the NOD allele is hypofunctional, that completely knocking out CD137 protects from diabetes. It is possible that both regulatory and effector mechanisms are affected in the CD137−/− mice; further studies may show a more profound cellular effect on Tregs when comparing CD137−/− Tregs to the CD137+ Treg subset. To fully understand these results, future studies using bone marrow chimeras, mixing studies using transfers employing specific subsets from normal NOD versus NOD CD137−/− mice, and cell-specific CD137 knockouts will be illuminating.

Finally, the results of Chen et al. (4) should be understood in the context of the complex role of CD137 in the immune system: not only does the knockout of Cd137 affect CD137 signaling on both T-cell effectors and Tregs, but the CD137 ligand (CD137L) also mediates “reverse” signaling and thus CD137L signaling is also changed in these mice (Fig. 1B) (14,15). CD137L signaling could be playing a major role in the NOD autoimmune process; CD137L knockout mice were recently shown to have significant protection from the murine equivalent of multiple sclerosis, experimental autoimmune encephalomyelitis (16). Finally, knockout of CD137 also deprives the immune system of soluble CD137, which probably mediates immunosuppressive effects (Fig. 1B) (13). Determining the effect of the CD137−/− on these immune subsystems and the role of these pathways in T1D will require many more experiments using this novel reagent.

Article Information

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


  • See accompanying original article, p. 68.

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. See for details.


| Table of Contents