strated that Mus81-Eme1 could cleave that most noncrossover recombinants References

Holliday junctions in a purified system. form by a mechanism that does not

Mutants lacking this enzyme had a
profound meiotic crossover defect
that could be rescued by expression
of a bacterial Holliday junction cleaving
enzyme. Cromie et al. (2006) now show
that mus81 mutants accumulate Hol-
liday junctions. Combined with earlier
observations in budding yeast (de los
Santos et al., 2003), the fission yeast
studies suggest that not only do the
predominant DNA intermediates dif-
fer between species, but the enzymes
most important for resolving those
intermediates may differ as well.
Differences in the recombination
mechanisms between fission and bud-
ding yeast argue against a reductionist
approach to describing the mecha-
nism, even within a single organism.
Indeed, a number of studies indicate
that not all recombination occurs via
the canonical double-strand break
repair mechanism, even in budding
yeast (Merker et al., 2003; Allers and
Lichten, 2001 and references therein).
Of particular note is work suggesting
include a Holliday junction intermediate
(Figure 1C) (Allers and Lichten, 2001).
Thus, it appears that individual mei-
otic recombination events proceed via
different mechanisms within an organ-
ism and that the prevalent mechanism
can differ between organisms. What
controls progression on one pathway
versus another? Does pathway preva-
lence vary from one locus to the next,
and, if so, why? Do different pathways
contribute distinct biological or evolu-
tionary functions? Can distinguishing
features of the various pathways give
us clues about mechanisms of regu-
lation? For example, are double Hol-
liday junctions essential for crossover
control? Answering these questions
requires development of better assays
for diagnostic features of each pathway
at a given locus as well as examination
of additional loci in each organism.
Finally, the new study emphasizes the
need to characterize recombination
intermediates in beasts other than
yeasts.
Allers, T., and Lichten, M. (2001). Cell 106,
47–57.
Bell, L.R., and Byers, B. (1983). Cold Spring
Harb. Symp. Quant. Biol. 47, 829–840.
Boddy, M.N., Gaillard, P.H., McDonald, W.H.,
Shanahan, P., Yates, J.R., 3rd, and Russell, P.
(2001). Cell 107, 537–548.
Cromie, G., Hyppa, R.W., Taylor, A.F.,
Zakharyevich, K., Hunter, N., and Smith, G.R.
(2006). Cell, this issue.
de los Santos, T., Hunter, N., Lee, C., Larkin,
B., Loidl, J., and Hollingsworth, N.M. (2003).
Genetics 164, 81–94.
Jessop, L., Allers, T., and Lichten, M. (2005).
Genetics 169, 1353–1367.
Merker, J.D., Dominska, M., and Petes, T.D.
(2003). Genetics 165, 47–63.
Schwacha, A., and Kleckner, N. (1995). Cell
83, 783–791.
Schwacha, A., and Kleckner, N. (1997). Cell
90, 1123–1135.
Szostak, J.W., Orr-Weaver, T.L., Roth-
stein, R.J., and Stahl, F.W. (1983). Cell 33,
25–35.

Sensory Neurons Link the Nervous

System and Autoimmune Diabetes
Helene Bour-Jordan1 and Jeffrey A. Bluestone2,*
1
UCSF Diabetes Center, Department of Medicine and Department of Pathology, University of California, San Francisco, 513
Parnassus Avenue, Box 0540, San Francisco, CA 94143, USA
2
UCSF Diabetes Center, HSW 1118, Box 0540, 513 Parnassus Avenue, San Francisco, CA 94143, USA
*Contact: jbluest@diabetes.ucsf.edu
DOI 10.1016/j.cell.2006.11.030

The initial factors that trigger the autoimmune response against pancreatic islets in the
nonobese diabetic (NOD) mouse are still unknown. In this issue of Cell, Razavi et al. (2006)
propose that a defect in a subset of sensory neurons innervating the pancreas plays a
major role in initiating the chain of events that will lead to local inflammation, islet destruc-
tion, and autoimmune diabetes.

Type 1 diabetes in humans is an destruction of the insulin-produc- resembling human type 1 diabetes
autoimmune disease in which T cells ing β cells. The nonobese diabetic (reviewed in Anderson and Blue-
target pancreatic islets of Langer- (NOD) mouse spontaneously devel- stone, 2005). Although both genetic
hans, leading to the progressive ops an autoimmune diabetic disorder and environmental factors contribute

Cell 127, December 15, 2006 ©2006 Elsevier Inc.  1097

to the development of autoim- by the nervous system can
mune diabetes, the precise eti- alter inflammation and insu-
ology and the initiating factors lin resistance and thus “indi-
that trigger the autoimmune rectly” affect the development
response against pancreatic of the autoimmunity. However,
islets are still unknown. In an equally plausible possibility
this issue, Razavi et al. (2006) is that an autoimmune com-
propose that a defect in sen- ponent targeting the nervous
sory neurons innervating the system directly influences the
pancreas contribute to insulin development of autoreactive
resistance and β cell stress in responses against pancreatic
NOD mice by initiating local islets. A high percentage of
inflammation and an autoim- diabetic individuals in Sardinia
mune attack on pancreatic and Canada have multiple
islets. These results suggest sclerosis—an autoimmune T
a clear link between autoim- cell-mediated demyelinating
munity, inflammation, and the process. Furthermore, Dosch
nervous system. However, is and colleagues reported T cell
the link between the nervous responses and autoantibodies
system and autoimmunity an directed at neural cell antigens
epiphenomenon, or is there a expressed in the central nerv-
direct link between autoanti- ous system or in the peripheral
genic determinants shared by nervous system in patients
these intertwined tissues? with type 1 diabetes (Winer et
Pancreatic islets are inner- al., 2001; Winer et al., 2003). In
vated by a network of primary fact, many proteins thought to
sensory afferent neurons, a play a role in type 1 diabetes
subset of which expresses have an expression pattern
the capsaicin receptor TRPV1 Figure 1. Linking Autoimmune Diabetes and the limited to the pancreas and
(transient receptor poten- Nervous System the nervous system. Finally,
tial vanilloid-1). Razavi et al. In normal mice (not shown), a feedback loop involving islet β autoimmune responses tar-
cells and sensory neurons expressing TRPV1 maintains bal-
(2006) describe two muta- anced levels of insulin and substance P. In NOD mice, insulin geting the nervous system
tions in conserved regions of secretion by islet β cells fails to properly stimulate the sensory have been described in the
the trpv1 gene in NOD mice neurons expressing TRPV1 to release neuropeptides due to NOD mouse. Blocking the
the presence of a hypofunctional polymorphism in the trpv1
and propose trpv1 as a can- gene. Suboptimal local levels of neuropeptides lead to insulin costimulatory CD28/B7-2 or
didate gene for the diabetes resistance and β cell stress as well as a local proinflammatory IL-2 pathways in NOD mice
susceptibility locus Idd4.1 on milieu, while physiological cell death of neurons and islet β resulted in the development
cells leads to the presentation of auto-antigens by profession-
chromosome 11. TRPV1 poly- al antigen-presenting cells (APCs) in draining lymph nodes. In-
of an autoimmune peripheral
morphism in NOD mice was filtration of Schwann cell and islet-specific T cells is sustained neuropathy (Salomon et al.,
associated with functional by the local proinflammatory milieu resulting from defective 2001; Setoguchi et al., 2005).
TRPV1 signals in sensory neurons.
defects in neurons express- In addition, T cell responses
ing TRPV1, including reduced and autoantibodies directed
secretion of substance P from pan- the authors does not fully account for at peri-islet Schwann cells and other
creatic nerve terminals. Importantly, the initiation of the autoimmune proc- neural structures in the pancreas were
restoration of this function in congenic ess in NOD mice. found in the NOD mouse. Moreover,
NOD mice expressing the B6 TRPV1 The multiple interactions between this autoimmune response targeted
allele or by local administration of neuronal and endocrine functions led neural antigens shared with islets
substance P led to reduced autore- to the concept of an integrated neu- such as glutamic acid decarboxylase
active T cell proliferation, islet infiltra- roendocrine system. Intriguingly, the (GAD65) as well as Schwann cell pro-
tion, and diabetes. Thus, the authors demonstration that insulin-producing teins such as S100β and glial fibril-
suggest that a molecular defect in the cells found in the brain of the fruit fly lary acidic protein (GFAP) (Winer et
pancreatic innervation system could Drosophila are functionally analogous al., 2001; Winer et al., 2003). Finally,
dramatically affect autoimmune dia- to pancreatic islet β cells led Rulifson Dosch and colleagues have argued
betes. However, previous studies et al. (2002) to hypothesize the exist- that responses against neural ele-
have shown that the Idd4 locus does ence of a common ancestral insu- ments play an important role in “initi-
not provide complete protection from lin-producing organ of neural origin. ating” autoreactive responses against
insulitis and diabetes, suggesting The Razavi et al. (2006) study raises islets. Winer et al. (2003) showed that
that the neuronal defect identified by the possibility that signals delivered the insulitis that precedes overt clini-

1098  Cell 127, December 15, 2006 ©2006 Elsevier Inc.

cal diabetes may be, in part, directed
against immature Schwann cells sur-
rounding pancreatic islets. Further-
more, T cell lines specific for GFAP
promote insulitis, whereas antigen-
based therapy inhibits diabetes.
Dosch and colleagues now extend
their model to propose that defects in
TRPV1+ neuron function, particularly
the reduced release of substance P,
are directly linked to the mild hyperin-
sulinemia and insulin resistance that
have been previously described in
young NOD mice (reviewed in Homo-
Delarche, 2004) and lead to β cell
stress. Surprisingly, the removal of
neurons expressing TRPV1 by neo-
natal capsaicin treatment resulted
in a similar outcome on anti-islet
autoimmunity as restoration of their
function. Razavi et al. (2006) explain
this paradox by a model in which only
subnormal local levels of neuropep-
tides are pathogenic and lead to β cell
stress and inflammation. Although
neuropeptides can have opposite
physiological effects at different con-
centrations, it will remain to be con-
firmed whether complete elimination
of neurons expressing TRPV1 affects
diabetes through distinct pathways,
such as antigen removal or increased
regulatory T cells as shown by Razavi
et al. (2006).
The authors propose that together
with reduced local levels of sub-
stance P, which has anti-inflamma-
tory effects, β cell stress contributes
to the recruitment of mononuclear
cells that marks the beginning of the
autoimmune process against NOD
islets (Razavi et al., 2006). Indeed,
insulin resistance was abrogated in
NOD-B6-Idd4 congenic mice or after
local administration of substance P
in NOD mice. This mechanism may
not be limited to substance P as the
transgenic expression in pancreatic
islets of calcitonin gene-related pep-
tide (CGRP), another neuropeptide
that has immunosuppressive proper-
ties and is released by sensory neu-
rons, resulted in reduced diabetes
incidence in NOD mice (Khachatryan
et al., 1997). From a kinetic stand-
point, it is remarkable that physiolog-
ical cell death of both neurons and
islet β cells occurs shortly after birth
in all strains of mice and may lead to
the presentation of autoantigens by
professional antigen-presenting cells
in draining lymph nodes. Thus, the
precise timing of antigen release and
presentation and of local inflammation
secondary to neuronal defects may
determine whether initial autoreactive
T cell responses are directed against
islet or neural antigens. Importantly,
the defective feedback loop between
neurons expressing TRPV1 and β
cells was observed in immunodefi-
cient NOD (NOD-SCID) mice as well,
supporting the hypothesis that it may
initiate local inflammation rather than
be a consequence of it. Furthermore,
there are patients with a complex
form of diabetes with unexplained
combinations of syndromes from
types 1 (autoimmunity) and 2 (insu-
lin resistance). These are sometimes
referred to as “type 1-1/2” and have
some characteristics consistent with
the syndromes described by Razavi
et al. (2006). Interestingly, abnormal β
cell function has also been described
in NOD mice and was believed to be
secondary to high levels of inflam-
matory mediators during the autoim-
mune process. However, recent
data by Chaparro et al. (2006) iden-
tify intrinsic alterations of pancreas
physiology in the absence of autoim-
mune responses in NOD-SCID mice,
including ER protein stress and insulin
resistance. The findings by Razavi et
al. (2006) possibly link these param-
eters commonly associated with
type 2 diabetes to abnormal sensory
neuron function in NOD mice. Con-
versely, insulin resistance and β cell
stress have been described in mouse
models that do not develop autoim-
Cell 127, December 15, 2006 ©2006 Elsevier Inc.  1099
mune diabetes. Thus, the findings
presented by Dosch and colleagues
ought to be put in the context of other
defects in central and peripheral tol-
erance afflicting the NOD mouse to
be included in a general model (Fig-
ure 1) integrating neuroendocrine
and immunological defects that lead
to autoimmune diabetes. Finally, this
report once again raises the question
of the tissue specificity of autoimmu-
nity in mice and humans. Particularly,
additional studies will be needed to
reconcile the fact that the principal
target of autoimmunity remains the
pancreatic islet in unmanipulated
NOD mice despite strong evidence of
the implication of the nervous system
in this process.
References
Anderson, M.S., and Bluestone, J.A. (2005).
Annu. Rev. Immunol. 23, 447–485.
Chaparro, R.J., Konigshofer, Y., Beilhack,
G.F., Shizuru, J.A., McDevitt, H.O., and Chien,
Y.H. (2006). Proc. Natl. Acad. Sci. USA 103,
12475–12480.
Homo-Delarche, F. (2004). ILAR J. 45, 237–258.
Khachatryan, A., Guerder, S., Palluault, F.,
Cote, G., Solimena, M., Valentijn, K., Millet, I.,
Flavell, R.A., and Vignery, A. (1997). J. Immu-
nol. 158, 1409–1416.
Razavi, R., Chan, Y., Afifiyan, F.N., Liu, X.J.,
Wan, X., Yantha, J., Tsui, H., Tang, L., Tsai, S.,
Santamaria, P., et al. (2006). Cell, this issue.
Rulifson, E.J., Kim, S.K., and Nusse, R. (2002).
Science 296, 1118–1120.
Salomon, B., Rhee, L., Bour-Jordan, H., Hsin,
H., Montag, A., Soliven, B., Arcella, J., Girvin,
A.M., Miller, S.D., and Bluestone, J.A. (2001).
J. Exp. Med. 194, 677–684.
Setoguchi, R., Hori, S., Takahashi, T., and Sak-
aguchi, S. (2005). J. Exp. Med. 201, 723–735.
Winer, S., Astsaturov, I., Cheung, R., Gunarat-
nam, L., Kubiak, V., Cortez, M.A., Moscarello,
M., O’Connor, P.W., McKerlie, C., Becker, D.J.,
and Dosch, H.M. (2001). J. Immunol. 166,
2831–2841.
Winer, S., Tsui, H., Lau, A., Song, A., Li, X.,
Cheung, R.K., Sampson, A., Afifiyan, F., Elford,
A., Jackowski, G., et al. (2003). Nat. Med. 9,
198–205.