Expression of Aire and the Early Wave of Apoptosis
Claudia E. Schaller,* Clifford L. Wang,* Gabriele Beck-Engeser,* Lindsie Goss,*
Hamish S. Scott,‡Mark S. Anderson,†and Matthias Wabl2*
Expression of the autoimmune regulator (Aire) protein in mice and humans is thought to be restricted to the medullary epithelial
and monocyte-dendritic cells of the thymus. There it mediates expression and presentation of a large variety of proteins, including
those that are peripheral organ-specific and are not expressed by other thymocytes. In this way, self-reactive T lymphocytes that
would attack peripheral cells producing these proteins are confronted with the self-Ags and, as a consequence, are deleted. In this
study, we show that Aire mRNA is also expressed in the testis—another tissue with promiscuous gene expression. Aire protein,
however, is expressed only sporadically in spermatogonia and spermatocytes. Transcription of genes that are under Aire control
in the thymus is unaffected by Aire in the testis. However, in mice with a disrupted Aire gene, the scheduled apoptotic wave of germ
cells, which is necessary for normal mature spermatogenesis, is reduced, and sporadic apoptosis in adults is increased. Because
Rag-1 deficiency does not abolish the effect, the adaptive immune system is not involved. We suggest that there is a link between
the scheduled and sporadic apoptotic processes and propose that scheduled apoptosis provides a counterselection mechanism that
keeps the germline stable. The Journal of Immunology, 2008, 180: 1338–1343.
endocrinopathy-candidiasis-ectodermal dystrophy (APECED) in
humans (3–6). In mice, Aire deficiency recapitulates many char-
acteristics of APECED (7–9), including reduced fertility (7). In the
thymus, Aire expression mediates expression and presentation of a
number of proteins that are otherwise expressed only in peripheral
tissues. Self-reactive T lymphocytes that would attack peripheral
cells producing these proteins are confronted with the self-Ags and
deleted (8, 10). Aire thus seems to be unique in that, by promis-
cuous gene expression in a specialized cell type, it helps to mimic
a transcriptome of multiple tissues (10). But it clearly also has a
role in presentation (only) of self-Ags, the expression of which is
not under its control (9, 11): Aire-deficient medullary epithelial
cells are less efficient presenters (11). At any rate, regardless of the
exact mechanism of Aire activity, in the thymus the end effect of
Aire expression is T cell apoptosis, which can be regarded a
“scheduled” event. This is different from “sporadic” apoptosis
events, which can be a consequence of, for example, viral infec-
tion, cell stress, or an “error catastrophe,” i.e., an accumulation of
he autoimmune regulator (Aire3protein) is thought to be
a transcription factor (1) and, perhaps, an E3 ubiquitin
ligase (2) that, when mutated, causes autoimmune poly-
numerous mistakes, in RNAs and/or proteins, severe enough to
prevent the cell from functioning.
Although the apoptosis for an adult cell that has lived its useful
life is easy to understand, scheduled apoptosis for immature cells
poses a greater conundrum. The conceptual framework of negative
selection in the thymus, for example, has taken decades to be
worked out by experiments, and the promiscuous gene expression
(10) mediated by Aire (8) as a basis for apoptosis to peripheral
self-Ags was recognized only recently. In this study, we investi-
gated Aire expression in the testis, another tissue with promiscuous
gene expression, and Aire’s role in the scheduled apoptosis in
Although sporadic apoptosis with no obvious function occurs
throughout the life of the normal adult testis, an early and massive
wave of germ cell apoptosis occurs in testis between 2 and 4 wk
after birth (12–14), with a peak after 3 wk. Spermatocytes (12, 13)
are thought to be affected most, with spermatogonia (14) also be-
ing affected. It is estimated that at least 80% of the germ cells are
eliminated in this early wave of scheduled apoptosis, during which
at any point in time slightly ?1% of cells stain as undergoing
apoptosis (12, 13). When apoptosis is impeded by forced expres-
sion of anti-apoptotic genes bcl2 or bclxL, normal spermatogenesis
is impaired; as a result, the affected mice are sterile (12, 15). The
causes and/or the pathways of apoptosis during the wave at 3 wk
appear to differ from those of sporadic apoptosis in adult mice.
Although apoptotic spermatocytes are found in adult testes, spo-
radic apoptosis is thought to occur mainly among spermatogonia
(12), and it is not decreased in the testes of anti-apoptotic bcl2 or
bclxL transgenic mice (12).
It has been proposed that the early wave of apoptosis is neces-
sary for the maintenance of a critical ratio of cells of some germ
cell stages to Sertoli cells (12, 13, 16). In this hypothesis, a pre-
sumed over-production of germ cells is viewed as a problem that
needs to be corrected via death of most of the cells over a relatively
short time period. But it has also been suggested that the early
wave is a consequence of mutated DNA, perhaps triggered in part
by incorrect DNA rearrangements during chromosomal crossing
*Department of Microbiology and Immunology,†Diabetes Center, University of Cal-
ifornia, San Francisco, CA 94143; and‡Division of Molecular Medicine, The Walter
and Eliza Hall Institute of Medical Research, Victoria, Australia
Received for publication July 10, 2007. Accepted for publication November 11, 2007.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1Supported by National Institutes of Health Grants R01 AI041570 and a BioSTAR
grant funded by the University of California and Medarex (to M.W.).
2Address correspondence and reprint requests to Dr. Matthias Wabl, Department of
Microbiology and Immunology, University of California, San Francisco, 513 Parnas-
sus Avenue, San Francisco, CA 94143. E-mail address: firstname.lastname@example.org
3Abbreviations used in this paper: Aire, autoimmune regulator; APECED, polyen-
Copyright © 2008 by The American Association of Immunologists, Inc. 0022-1767/08/$2.00
The Journal of Immunology
over in the first meiotic division (12, 13). These hypotheses are not
mutually exclusive, but the relative contribution of each process to
apoptosis is not known. However, the second hypothesis—mutated
DNA—provides a potential connection to the sporadic apoptosis
that occurs later: if mutant germ cells are not purged, the effects of
mutations would become evident later in life, and among these
effects would be apoptosis.
In this study, we first investigate the effect of Aire expression on
both scheduled and sporadic apoptosis in the testis. In the absence
of a consensus on what causes these events in the first place, we
also thought it useful to address this issue independently from Aire
expression. In this way, the effect of Aire may be interpreted in a
more cogent way. Thus we investigated whether we could disturb
the critical ratio of germ cells to Sertoli cells and still maintain
normal fertility. In that case, the mutated DNA hypothesis would
be favored over the critical ratio hypothesis. Mice with a defi-
ciency in the DNA mismatch repair gene, pms2 gene, were used in
these experiments. Wild-type pms2 not only contributes to
genomic integrity through DNA repair but also has been linked to
the apoptosis (17) of cells damaged beyond repair (18). Because
somatic pms2 knockout mouse cells apoptose less in response to
DNA damage (19), we thought that pms2-deficient germ cells may
also do so. Unlike the male mice that have pms2 knocked out (20),
males with a dominant-negative mutant transgene of the mismatch
repair protein pms2 (called morphogene) are fertile (21) and thus
could be used for our experiment. As expected, their progeny also
carry an increased mutational load (21) that may or may not be a
consequence of the putative reduction in germ cell apoptosis.
Materials and Methods
In this study, we have used two independently derived Aire-deficient mice
(LA, Aire-deficient mouse strain described by Ramsey et al. (7); HD, Aire-
deficient mouse strain described by Anderson et al. (8)), which for the
purpose of this paper we have named HD Aire?/? and LA Aire?/?. (In
the following, to distinguish between the origins of the mice, whether HD
or LA derived, we add the prefix HD or LA to both Aire-deficient and
Aire-sufficient mice.) In the HD Aire-deficient mouse, exon 2 of Aire is
deleted by cre-mediated recombination (8). In the LA Aire-deficient
mouse, Aire exons 5 and 6 are replaced by the neomycin resistance gene
(7). When the HD mice arrived at our mouse house, they were on the
C57BL/6 background, and we kept backcrossing them to that strain. The
LA mice were not on a homogenous background, as immediately evident
by the different coat colors of the offspring; we began backcrossing them
to C57BL/6; however, at the time when the experiments were performed,
there were not enough generations to consider them backcrossed. The male
Aire-deficient mice were less fertile than their Aire-sufficient counterparts.
Therefore, to obtain the mice described in the experiments, we always used
heterozygous males (Aire?/?); females were either Aire?/? or Aire?/?.
The Rag-1-deficient mice were on a C57BL/6 background. In the experi-
ments with Rag mice, HD Aire?/? Rag?/? were compared with HD
Aire?/? Rag?/?. The Morphomouse (21) was on an undefined back-
ground; it contains the so-called morphogene, which is a dominant mutant
transgene of the mismatch repair protein pms2. The Morphomouse was
bred to HD and LA mice, respectively.
Preparation of tissue sections
For TUNEL assay and histological analysis, testes were dissected and fixed
in freshly made Bouin’s solution for 8 to 10 h at room temperature. Testes
were washed several times in 70% ethanol and embedded in paraffin ac-
cording to standard procedures. Embedded testes were sectioned (4 ?m)
using Fisherbrand Superfrost microscope slides (Thermo Fisher Scientific);
sections were cut out of five tissue levels. Slides were stored at room
temperature in the dark until further processing. For immunohistochemical
staining, testes were frozen in Tissue-Tek O.C.T. compound (Electron Mi-
croscopy Sciences) using dry ice covered with 2-methyl butane. Frozen
sections (4 ?m) were cut out of five tissue levels, mounted on Fisherbrand
Superfrost microscope slides, and stored at ?80°C until further processing.
Histological analysis was conducted using Harris hematoxylin (Surgipath)
for nuclear staining and eosin (Surgipath) for cytoplasmic staining.
DNA fragments were labeled using the Roche in situ cell death detection
kit, POD (Roche Diagnostics). After deparaffinization in fresh xylene,
slides were rehydrated in a graded series of ethanol diluted in double-
distilled water. Samples were immersed in PBS followed by 3% H2O2in
methanol for 10 min at room temperature to inactivate endogenous perox-
idase activity. After washing in PBS, the tissue was permeabilized in 0.1 M
citrate buffer (pH 6.0), with microwave irradiation at 500 W for 70 s, and
subsequently transferred into PBS. Sections were blocked in 3% BSA in
PBS for 25 min at room temperature. TUNEL reaction mixture containing
label solution (Roche) with nucleotide mixture and enzyme solution
(Roche) with TdT was prepared according to the manufacturer’s guidelines
and applied. Slides were incubated at 37°C for 60 min in a humidified
chamber. After washing in PBS, converter peroxidase containing anti-
fluorescein Ab conjugated with HRP was applied for 30 min at 37°C in a
humidified chamber. 3,3?-diaminobenzidine was used as a peroxidase sub-
strate (Vector Laboratories) intensified with Ni2?to obtain dark brown
staining. Sections were counterstained with Nuclear Fast Red (Vector Lab-
oratories) for 10 min, dehydrated in ethanol, cleared in xylene, mounted,
For positive labeling control, sections were treated with DNase I,
grade I (1500 U/ml in 50 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 1
mg/ml BSA) for 30 min at 37°C in a humidified chamber to induce
DNA strand breaks. For negative labeling control, sections were incu-
bated in label solution only.
Frozen sections were air-dried for 30 min, fixed in ice-cold acetone for 10
min and air-dried again. Samples were washed in PBS followed by im-
mersion in 0.5% H2O2in PBS for 10 min at room temperature to inactivate
endogenous peroxidase activity. Tissue was blocked in 10% (v/v) normal
rabbit serum in PBS for 30 min at room temperature in a humidified cham-
ber. Subsequently, the sections were incubated with monoclonal rat anti-
Aire Ab (clone 5H12, IgG2c isotype) in PBS with 3% BSA for 60 min at
room temperature. After washing in PBS, mouse Ig absorbed, biotinylated
rabbit anti-rat secondary Ab (Vector Laboratories) in PBS with 3% BSA
was applied for 45 min at room temperature in a humidified chamber. After
immersion in PBS, tissue was incubated in streptavidin-HRP in PBS for 30
min at room temperature. The substrate for peroxidase was 3,3?-diamino-
benzidine (Vector Laboratories); counterstaining was done with Nuclear
Fast Red (Vector Laboratories) for 5 min, and slides were mounted and
Images were viewed with a light microscope (Leica, ?40) and recorded
with a digital camera. A germ cell was considered apoptotic when it
showed dark brown intense nuclear staining. Apoptotic germ cells in 20
randomly focused seminiferous tubules per level (100 tubules per testis)
were counted and divided by the total number of all germ cells in all
evaluated seminiferous tubules, then multiplied by 100 to obtain the per-
centage of apoptotic cells. Germ cell stages of apoptotic cells were iden-
tified according to their morphology and position in the seminiferous tu-
bule. Aire-positive cells were counted in 20 randomly focused
seminiferous tubules per level (100 tubules per testis) and divided by the
total, average number of all germ cells, multiplied by 100, to obtain the
percentage of Aire-positive cells.
Quantitative PCR analysis
Gene expression was measured using quantitative RT-PCR. RNA was ex-
tracted from cells, and cDNA was produced by standard application of
reverse transcriptase and oligo-dT primers. Each mouse used in the study
is shown in Table I. In the 23-day-old testes, each group contained 2 mice.
In the adult testes, each group contained 3 mice. For each mouse, the
quantitative RT-PCR reactions were performed in triplicate. PCR was per-
formed using the ABI PRISM 7700 machine. Primers and probes were as
AIRE: 5?-GCCAAGGGAGCCCAGG-3?, 5?-GGAGGAACCCCACA
GTGGG-(TAMRA)-3?, Hbb-y: 5?-GCTAGTCACTTCGGCAATGAATT-
3?, 5?-CCCCAGCCACCAGCTTC-3?, 5?-(6-FAM)- AGCTGAGATGCA
GGCTGCCTGGC-(TAMRA)-3?; NPY: 5?-CAGAAAACGCCCCCAGA
AC-3?, 5?-CGGGAGAACAAGTTTCATTTCC-3?, 5?-(6-FAM)-AGGCTT
1339The Journal of Immunology
GAAGACCCTTCCATGTGGTGA-(TAMRA)-3?; Ptdgs: 5?-CCTGCCCC
AACCGGAT-3?, 5?-GTGACCAGCCCTCTGACTGAC-3?, 5?-(6-FAM)-
ACA-(TAMRA)-3?; Spt2: 5?-CACCATGAAGTTCCTGGCACT-3?, 5?-
TCCACTATCCTAGTC-(TAMRA)-3?; IRBP: 5?-AATGACTCGGTCAG
CGAACTTT-3?, 5?-CTGTCACACCACTGGTCAGGAT-3?, 5?-(6-FAM)-
ACAGGTGAACGATATGGCTCCAAG AAG-(TAMRA)-3?; FABP: 5?-
(TAMRA)-3?; proinsulin: 5?-ATCTACAATGCCACGCTTCTG-3?, 5?-
ACCTT-(TAMRA)-3?; ?-actin: 5?-AGGTCATCACTATTGGCAACGA-
3?, 5?-CACTTCATGATGGAATTGAATGTAGTT-3?, 5?-(6-FAM)-TG
Aire mRNA and protein expression in the testis
Reports on the tissue expression of Aire vary considerably. Aire is
said to be somewhat widely distributed (22) or to be restricted to
the peripheral monocyte/dendritic cell lineage (23) in thymus (8,
24–26) and lymph nodes (26), or to ovary and testis (8, 25). It is
also not clear 1) whether the mRNA actually directs translation
into Aire protein in ovary and testis, 2) in which cell type it is
expressed, and 3) if the message is indeed translated, whether the
protein has a functional role in spermatogenesis. One reason for
the discrepancies might be the use of polyclonal Abs. Using a
mAb, we found Aire protein expressed in the medullary cells of the
thymus, in spermatogonia, and in early spermatocytes, as differ-
entiated by position within the tubule and their morphology (Fig.
1). In thymus, only medullary cells stained (Fig. 1C), and there
was no staining with irrelevant Ab (isotype control) (Fig. 1A) or in
tissue from Aire-deficient mice (Fig. 1B). In testis, spermatogonia
and spermatocytes stained, but not spermatids or Sertoli cells (Fig.
1F), and no staining was observed in Aire-deficient tissue (Fig. 1E)
or with irrelevant Ab (isotype control) (Fig. 1D). We then counted
the cells in 3-wk- and 3-mo-old mice. To mitigate potential
founder effects of a knockout line, we used two Aire-deficient
mouse strains that were independently generated in two different
laboratories, here designated LA Aire?/? (7) and HD Aire?/?
(8). HD was on the C57BL/6 strain, whereas LA had unknown
strain contributions. In the following, to distinguish between the
origins of the mice, whether HD or LA derived, we add the prefix
HD or LA to Aire-deficient or Aire-sufficient mice.
In testis tissue from 3-wk-old Aire-sufficient littermates of LA
Aire?/? and HD Aire?/? mice, ?0.4% (0.43 and 0.41%, re-
spectively) of cells were Aire positive, and in 3-mo-old mice,
Immunohistochemical stains on fixed frozen sections using mAb to mouse
Aire protein. A–C, Medullary region of thymus; D–F, tubules of testis. A
and D, with isotype control; B and E, with anti-Aire Ab on Aire-deficient
tissue; C and F, with anti-Aire Ab on Aire-sufficient tissue. ?/?, Aire
wild-type tissue; ?/?, Aire deficient tissue. Arrows point to Aire-positive
germ cells in testis. G, Percentage of Aire-positive cells in testis tissue of
3-wk- and 3-mo-old Aire-sufficient littermates (Aire?/?) from breedings
of LA and HD mice. Each data set is from five mice, 100 tubules each. H,
TaqMan analysis of Aire mRNA expression in testicular cell lysates, from
Aire-deficient and Aire-sufficient HD mice, three mice each. Left, Testis of
3-wk- (dark) and 3-mo-old (light) mice; right, thymus. Data normalized to
actin expression. The error bar attached to the value of the 3-wk-old mice
is very small and thus hardly visible.
Aire protein is expressed in the germ cells of the testis.
Table I. mRNA expression in mouse thymus and testis relative to ?-actin mRNAa
aMice of the same age were siblings.
bUnd, Undetected expression.
1340AIRE EXPRESSION IN TESTIS
0.15% of cells were Aire positive (Fig. 1G). The intense staining
suggests that a substantial quantity of Aire protein was present,
consistent with the results from quantitative RT-PCR: the ratio of
Aire to ?-actin expression was between 0.8 ? 10?4(3-wk-old)
and 2.2 ? 10?4(3-mo-old) (Fig. 1H). Considering that only 1 in
240 (3-wk-old) and 1 in 670 (3-mo-old) germ cells express Aire,
but every cell expresses actin, we multiply by 240, or 670, to
obtain a range of the Aire-to-?-actin ratio per cell. Accordingly,
we find that Aire is highly expressed in these cells—only 10- to
35-fold less abundant than actin; however, this is still an un-
derestimate, because we did not consider the presence of non-
germ cells (e.g., Sertoli cells) in this analysis. We note, how-
ever, that there is no obvious defined differentiation state in
which all the germ cells express Aire, implying sporadic ex-
pression. It is widely known that in testis one can find expressed
sequence tags for numerous genes that have specific functions in
other tissues; this promiscuous mRNA expression has been viewed
as an inconsequential side effect of the removal of the epigenetic
marks on the genome, which is necessary to “reset” the develop-
mental program. Along this line, Aire mRNA might simply be a
passenger; in which case its translated product would be without
Genes regulated by Aire in thymus under different control in
Because the promiscuous gene expression in the testis resembles,
at least superficially, that in the thymus, we tested whether Aire
expression also contributes to it. We picked four genes known to
be under Aire control in the thymus and assayed their expression
by quantitative RT-PCR in 23-day-old thymus and testis; we also
tested a total of seven genes in adult testis (Table I). Fig. 2 shows
the ratio of expression of these genes in Aire heterozygous over
Aire-deficient mice, as an indicator of Aire-dependent gene ex-
pression—“no effect” would read as 1.0. In the thymus (Fig. 2A),
this ratio was well above 1.0 for all five genes (please note the
logarithmic scale). This is notably high, as the signal was diluted
by the 99% of thymic cells that do not express Aire, and because
there is a substantial gene dosage effect of Aire in heterozygotes
(27). Because there are even fewer Aire-expressing cells in the
testis, signal dilution will be even greater. Nevertheless, it seems
clear that in the 23-day-old testis (Fig. 2B), the four genes were not
influenced by Aire. In adult testis (Fig. 2C), there was also no
indication for positive Aire control, with the possible exception of
Spt2, which we did not test in the 23-day-old testis. If anything,
there might have been some suppression. This means that the sub-
set of genes studied here, although promiscuously expressed in
both thymus and testis, is under different transcriptional control.
Similarly, a transgene encoding hen egg lysozyme under the in-
sulin promoter was expressed in the thymus of Aire-sufficient but
not Aire-deficient mice; however, in the testis it was expressed in
both types of mice (27).
Scheduled and sporadic apoptosis in Aire-deficient mice
One of the dramatic events in spermatogenesis is scheduled ap-
optosis. Vaguely reminiscent of counterselection of T cells in the
thymus, fewer than 20% of germ cells survive in this early apop-
totic wave (12, 13). We tested whether Aire has a role in the
apoptotic wave of 3-wk-old mice. In the seminiferous tubules we
counted (100 tubules per mouse, 5 mice) the apoptotic cells using
the TUNEL assay, and compared percentages of apoptosed germ
cells over all germ cells between Aire-sufficient and the two types
of Aire knockout mice, LA and HD. In Aire-sufficient (Aire?/?)
mice, on average we scored 1.6% apoptotic cells (Fig. 3), which is
in good agreement with the average 1.4% reported previously (12),
measured by the same technique. Counting all the germ cells in the
seminiferous tubules, regardless of their differentiation stage, in
both HD and LA Aire-deficient mice, there were only 1.2% apop-
totic cells, which represents a 25% reduction as compared with
Aire-sufficient siblings (p ? 0.001 for LA mice; p ? 0.0005 for
HA mice) (Fig. 3). This reduction in apoptosis was similar in mice
on a Rag-1-minus background (Fig. 3), indicating that it is inde-
pendent of the adaptive immune system. The reduction was more
pronounced when we differentiated between cell types: the sper-
matocytes II were affected the most, with only half as many ap-
optotic cells in LA (p ? 0.08) and HD (p ? 0.0003) Aire-deficient
mice, compared with Aire-sufficient (Aire?/?) (Fig. 4).
The causes and/or pathways of apoptosis during the scheduled
wave at 3 wk seem to differ from those of sporadic apoptosis in
adult mice. Unlike scheduled apoptosis, sporadic apoptosis is not
decreased in the testes of anti-apoptotic bcl2 or bclxL transgenic
mice (12); if anything, there is an increase (wt, 0.21 ? 0.09; bclxL
transgenic, 0.39 ? 0.1) (12). Similar to the findings reported by
Rodriguez et al. (12), we found that apoptosis in 3-mo-old Aire-
sufficient mice dropped substantially in our study to 0.3%, i.e., to
one fifth of the 3-wk level (Fig. 3). However, in contrast to the
adult testis of the anti-apoptotic bclxL transgenic mice, which have
highly disorganized tubules, we found no anatomic abnormalities
in Aire-deficient tubules in 3-mo-old mice. But the number of ap-
optotic germ cells increased by a factor of 2, from ?0.3 to 0.6%
(p ? 0.00005 for LA mice; p ? 0.00001 for HA mice) (Fig. 3). All
thymus; B, 23-day-old testis; C, adult testis. AIRE dependence is repre-
sented as the quotient of gene expression levels in AIRE heterozygous and
deficient mice. FABP, fatty acid binding protein 2 intestinal; Proins, pro-
insulin; Ptdgs, prostaglandin-D2 synthase; Spt1, salivary protein 1; IRBP
retinol-binding protein 3 interstitial; Hbb-y, hemoglobin Y ?-like embry-
onic chain; NPY, neuropeptide Y; Spt2, salivary protein 2 precursor. FABP
was undetected in the 23-day testes, and IRBP was undetected in the thy-
mus. The tests on their expression were performed but there was no inter-
pretable signal in these after 40 cycles.
Dependence of gene expression on AIRE. A, 23-day-old
1341 The Journal of Immunology
germ cell stages seemed to be affected, from spermatogonia
through spermatocytes to later stages (Fig. 4).
Scheduled apoptosis in mismatch repair-deficient mice
Left with the two opposite effects in spermatogenesis of 3-wk- and
3-mo-old Aire-deficient mice, we wondered about the underlying
causes of apoptosis in Aire-sufficient mice. As mentioned above, it
has been proposed that the early wave is necessary for the main-
tenance of a critical ratio of cells of some germ cell stages to
Sertoli cells (12, 13). Alternatively, the effects of mutated DNA
could trigger the apoptosis (12, 13). According to the first hypoth-
esis, if the apoptotic wave were necessary solely to maintain the
critical germ cell number to satisfy a rather strict stoichiometry,
almost any substantial reduction in apoptotic cells ought to disrupt
normal spermatogenesis, regardless of the underlying cause of that
reduction. According to the second hypothesis, however, a forced
reduction in apoptosis would not necessarily have as dramatic an
effect on spermatogenesis as, for example, the bcl2 transgene has,
but instead would increase the number of cells with mutations.
To address these issues, we investigated the effect of a dominant
negative mutant transgene of the mismatch repair protein pms2.
This mutant gene, referred to as a morphogene (21), predisposes
humans to hereditary nonpolyposis colon cancer (28). When in-
troduced into the cells of bacteria, yeasts, plants and mammalian
cells, it increases the rate of genome-wide mutagenesis (21). Be-
cause pms2 stabilizes the apoptosis-activating protein p73 (17) and
is a direct target of p53 (18), it not only contributes to genomic
integrity through DNA repair, but is also a link to apoptosis of cells
damaged beyond repair, in this way reducing the mutational load.
Because somatic pms2 knockout mouse cells apoptose less in re-
sponse to DNA damage (19), we thought that transgenic morpho-
gene germ cells may also do so. Indeed, this is the case: In 3-wk-
old mice the percentage of apoptotic cells was reduced to almost
half that in wt (i.e., no morphogene) (0.9 and 1.6, respectively; p ?
0.0001) (Fig. 3). Nevertheless, there were no anatomic abnormal-
ities in the tubules of either the 3-wk-old or the 3-mo-old mice.
Furthermore, the apoptotic indices of 3-mo-old mice did not
differ significantly from those of wt (Fig. 3). Clearly, the morpho-
gene reduces apoptosis and thus can interfere with the ratio of
germ cells to Sertoli cells. Because this does not affect adult sper-
matogenesis but increases mutational load, we think that this is
evidence for a contribution postulated by the second hypothesis,
which interprets the apoptotic wave as a consequence of mutated
DNA. If this is correct, then up to 80% of the germ cells in 3-wk-
old Aire-sufficient mice would contain deleterious mutations; as
discussed above, at this age over 80% of cells are deleted, which
translates into the 1.6% steady-state level of apoptotic cells scored
in the tissue sections.
In this paper, we report on our studies of the early wave of sched-
uled apoptosis in spermatogenesis in Aire-deficient and mismatch
repair-deficient mice. Because normal spermatogenesis requires
elaborate interactions between Sertoli and germ cells and because
the apoptotic wave spares the Sertoli cells, it was proposed that the
ratio of the different stages of germ cells to Sertoli cells is critical
and that a disruption of this process would permanently impair
spermatogenesis (12). We have shown here that the reduced ap-
optosis in 3-wk-old morphogene mice has no obvious effect on the
morphology and further differentiation of germ cells. This leads us
to conclude that scheduled apoptosis is not a developmental ne-
cessity to maintain the definite proportions between Sertoli and
germ cells. Rather, we suggest that it is a checkpoint of genomic
health. Germ cells that fail to complete correct DNA rearrange-
ments during the chromosomal crossing over of the pachytene
phase of the first meiotic division, for example, will apoptose (12).
However, we also found that reduced apoptosis in the 3-wk-old
Aire-deficient mice is correlated with increased sporadic apoptosis
at 3 mo. This long-range effect may be explained by suggesting
that some of the mutations introduced up to the 3-wk checkpoint
would render the cells nonfunctional only later, such as at 3 mo.
Although we do not know the underlying cause for the sporadic
apoptosis, we think it reasonable to assume that cells poised to
proliferate and further differentiate, such as spermatocytes, die as
a consequence of deleterious mutations.
cell differentiation stage. A and B, HD; C and D, LA. Each data set is from
5 mice, 100 tubules each; 3 wk, testis from 3-wk-old mice; 3 mo, testis
from 3-mo-old mice; Spg, spermatogonia; Spc I, spermatocytes I; Spc II,
spermatocytes II; eSpd, early spermatids; Spd, spermatids.
Percentage of apoptotic germ cells, stratified according to
represent the arithmetic mean with SD. Each data set is from 5 mice (ex-
cept for two HD Aire?/?, Rag-1?/? mice), 100 tubules each. (In some
experiments, we also included a small number of Aire?/? mice, with
numbers of apoptotic cells no different from Aire?/? mice). 3 wk, testis
from 3-wk-old mice; 3 mo, testis from 3-mo-old mice; morphogene ?,
mice with the dominant-negative mismatch repair morphogene. Similar to pa-
tients with APECED, who often suffer from sterility associated with testicular
atrophy (6), Aire-deficient mice have reduced fertility, with no anatomic ab-
normalities and no lymphocyte infiltration apparent in the testes (7, 9). Nev-
ertheless, the HD mice on a Rag-1-deficient background bred well.
Percentage of apoptotic (TUNEL positive) germ cells. Bars
1342 AIRE EXPRESSION IN TESTIS
The reduction in scheduled apoptosis in the Aire-deficient mice
is highly reproducible. Because Aire-deficient mice on a Rag-1-
minus background also show the effect, it is not due to the adaptive
immune system. However, its attribution to the missing Aire pro-
tein is unclear. In the mouse, tightly linked to and partially over-
lapping with the Aire locus is the Dnmt3L gene, transcribed in the
opposite direction. Dnmt3L is a member of the DNA methyltrans-
ferase 3 family that lacks enzymatic activity but is required for de
novo methylation of imprinted genes in oocytes (29) and for trans-
poson repression in male germ cells (30). Although the exons of
Aire and Dnmt3L do not overlap, 5 kb from the Aire promoter, a
promoter active in prospermatogonia drives transcription of an
mRNA encoding the full-length Dnmt3L protein in perinatal testis
(31), where de novo methylation occurs. Late pachytene spermato-
cytes activate a second promoter in intron 9 of the Dnmt3L gene,
approximately 11 kb from the Aire promoter.
Loss of Dnmt3L from early germ cells causes meiotic failure
(“meiotic catastrophe”) in spermatocytes, which themselves do not
express Dnmt3L (30). Although in the Aire-deficient mice, the
exons of the Dnmt3L genes are untouched, the regions upstream of
the promoters active in testis are affected. In the HD mouse, exon
2 of Aire, along with a small DNA segment of the first intron of the
oocyte Dnmt3L transcript, is deleted by cre-mediated recombina-
tion (8); but the deletion leaves intact the prospermatogonia and
spermatocyte/spermatid transcript of Dnmt3L. In the LA mouse,
Aire exons 5 and 6 are replaced by the neomycin resistance gene,
approximately 1.7 kb upstream from the Dnmt3L oocyte promoter
and 6.7 kb upstream from the prospermatogonia promoter. It is
conceivable that in both HD and LA mice, disruption of sequence
elements upstream from the Dnmt3L promoter leads to hypomorph
phenotypes. But independent of this, and in regard to the open
question of which of the genes is important for the scheduled
apoptosis, our data suggest that there is a checkpoint for counter-
selection of germ cells with mutant genes. We speculate that the
promiscuous gene expression found in testis may serve as quality
control: Cells with mutated genes would apoptose, perhaps as a
consequence of the unfolded protein response. Cells that cannot
apoptose, as in the morphogene mouse, would carry a higher
We thank Diane Mathis and Leena Peltonen for Aire-deficient mice; Nick
Nicolaides for the morphogene mouse; Pa ¨rt Peterson for Abs; Nils Lon-
berg, Mindy Brooks, Nick Salmon and Renee Reijo Pera for discussions
and support; and Margaret Mayes for technical help. This work was per-
formed by C.S. as part of the requirement for the Ph.D. degree at the
The authors have no financial conflict of interest.
1. Kumar, P. G., M. Laloraya, C. Y. Wang, Q. G. Ruan, A. Davoodi-Semiromi,
K. J. Kao, and J. X. She. 2001. The autoimmune regulator (AIRE) is a DNA-
binding protein. J. Biol. Chem. 276: 41357–41364.
2. Uchida, D., S. Hatakeyama, A. Matsushima, H. Han, S. Ishido, H. Hotta,
J. Kudoh, N. Shimizu, V. Doucas, K. I. Nakayama, et al. 2004. AIRE functions
as an E3 ubiquitin ligase. J. Exp. Med. 199: 167–172.
3. Nagamine, K., P. Peterson, H. S. Scott, J. Kudoh, S. Minoshima, M. Heino,
K. J. Krohn, M. D. Lalioti, P. E. Mullis, S. E. Antonarakis, et al. 1997. Positional
cloning of the APECED gene. Nat. Genet. 17: 393–398.
4. Finnish-German APECEDConsortium.
APECED, caused by mutations in a novel gene featuring two PHD-type zinc-
finger domains. Nat. Genet. 17: 399–403.
5. Aaltonen, J., P. Bjorses, L. Sandkuijl, J. Perheentupa, and L. Peltonen. 1994. An
autosomal locus causing autoimmune disease: autoimmune polyglandular disease
type I assigned to chromosome 21. Nat. Genet. 8: 83–87.
6. Bjorses, P., J. Aaltonen, N. Horelli-Kuitunen, M. L. Yaspo, and L. Peltonen.
1998. Gene defect behind APECED: a new clue to autoimmunity. Hum. Mol.
Genet. 7: 1547–1553.
7. Ramsey, C., O. Winqvist, L. Puhakka, M. Halonen, A. Moro, O. Kampe,
P. Eskelin, M. Pelto-Huikko, and L. Peltonen. 2002. Aire deficient mice develop
multiple features of APECED phenotype and show altered immune response.
Hum. Mol. Genet. 11: 397–409.
8. Anderson, M. S., E. S. Venanzi, L. Klein, Z. Chen, S. P. Berzins, S. J. Turley,
H. von Boehmer, R. Bronson, A. Dierich, C. Benoist, and D. Mathis. 2002.
Projection of an immunological self shadow within the thymus by the aire pro-
tein. Science 298: 1395–1401.
9. Kuroda, N., T. Mitani, N. Takeda, N. Ishimaru, R. Arakaki, Y. Hayashi,
Y. Bando, K. Izumi, T. Takahashi, T. Nomura, et al. 2005. Development of
autoimmunity against transcriptionally unrepressed target antigen in the thymus
of Aire-deficient mice. J. Immunol. 174: 1862–1870.
10. Derbinski, J., A. Schulte, B. Kyewski, and L. Klein. 2001. Promiscuous gene
expression in medullary thymic epithelial cells mirrors the peripheral self. Nat.
Immunol. 2: 1032–1039.
11. Anderson, M. S., E. S. Venanzi, Z. Chen, S. P. Berzins, C. Benoist, and
D. Mathis. 2005. The cellular mechanism of Aire control of T cell tolerance.
Immunity 23: 227–239.
12. Rodriguez, I., C. Ody, K. Araki, I. Garcia, and P. Vassalli. 1997. An early and
massive wave of germinal cell apoptosis is required for the development of func-
tional spermatogenesis. EMBO J. 16: 2262–2270.
13. Mori, C., N. Nakamura, D. J. Dix, M. Fujioka, S. Nakagawa, K. Shiota, and
E. M. Eddy. 1997. Morphological analysis of germ cell apoptosis during postnatal
testis development in normal and Hsp 70–2 knockout mice. Dev. Dyn. 208:
14. Wang, R. A., P. K. Nakane, and T. Koji. 1998. Autonomous cell death of mouse
male germ cells during fetal and postnatal period. Biol. Reprod. 58: 1250–1256.
15. Furuchi, T., K. Masuko, Y. Nishimune, M. Obinata, and Y. Matsui. 1996. Inhi-
bition of testicular germ cell apoptosis and differentiation in mice misexpressing
Bcl-2 in spermatogonia. Development 122: 1703–1709.
16. Print, C. G., and K. L. Loveland. 2000. Germ cell suicide: new insights into
apoptosis during spermatogenesis. BioEssays 22: 423–430.
17. Shimodaira, H., A. Yoshioka-Yamashita, R. D. Kolodner, and J. Y. Wang. 2003.
Interaction of mismatch repair protein PMS2 and the p53-related transcription
factor p73 in apoptosis response to cisplatin. Proc. Natl. Acad. Sci. USA 100:
18. Chen, J., and I. Sadowski. 2005. Identification of the mismatch repair genes
PMS2 and MLH1 as p53 target genes by using serial analysis of binding ele-
ments. Proc. Natl. Acad. Sci. USA 102: 4813–4818.
19. Zeng, M., L. Narayanan, X. S. Xu, T. A. Prolla, R. M. Liskay, and P. M. Glazer.
2000. Ionizing radiation-induced apoptosis via separate Pms2- and p53-depen-
dent pathways. Cancer Res. 60: 4889–4893.
20. Baker, S. M., C. E. Bronner, L. Zhang, A. W. Plug, M. Robatzek, G. Warren,
E. A. Elliott, J. Yu, T. Ashley, N. Arnheim, et al. 1995. Male mice defective in
the DNA mismatch repair gene PMS2 exhibit abnormal chromosome synapsis in
meiosis. Cell 82: 309–319.
21. Nicolaides, N. C., W. Ebel, B. Kline, Q. Chao, E. Routhier, P. M. Sass, and
L. Grasso. 2005. Morphogenics as a tool for target discovery and drug develop-
ment. Ann. NY Acad. Sci. 1059: 86–96.
22. Halonen, M., M. Pelto-Huikko, P. Eskelin, L. Peltonen, I. Ulmanen, and
M. Kolmer. 2001. Subcellular location and expression pattern of autoimmune
regulator (Aire), the mouse orthologue for human gene defective in autoimmune
polyendocrinopathy candidiasis ectodermal dystrophy (APECED). J. Histochem.
Cytochem. 49: 197–208.
23. Kogawa, K., S. Nagafuchi, H. Katsuta, J. Kudoh, S. Tamiya, Y. Sakai,
N. Shimizu, and M. Harada. 2002. Expression of AIRE gene in peripheral mono-
cyte/dendritic cell lineage. Immunol. Lett. 80: 195–198.
24. Heino, M., P. Peterson, J. Kudoh, K. Nagamine, A. Lagerstedt, V. Ovod,
A. Ranki, I. Rantala, M. Nieminen, J. Tuukkanen, et al. 1999. Autoimmune
regulator is expressed in the cells regulating immune tolerance in thymus me-
dulla. Biochem. Biophys. Res. Commun. 257: 821–825.
25. Heino, M., P. Peterson, N. Sillanpaa, S. Guerin, L. Wu, G. Anderson, H. S. Scott,
S. E. Antonarakis, J. Kudoh, N. Shimizu, et al. 2000. RNA and protein expression
of the murine autoimmune regulator gene (Aire) in normal, RelB-deficient and in
NOD mouse. Eur. J. Immunol. 30: 1884–1893.
26. Klamp, T., U. Sahin, B. Kyewski, J. Schwendemann, K. Dhaene, and O. Tureci.
2006. Expression profiling of autoimmune regulator AIRE mRNA in a compre-
hensive set of human normal and neoplastic tissues. Immunol. Lett. 106: 172–179.
27. Liston, A., D. H. Gray, S. Lesage, A. L. Fletcher, J. Wilson, K. E. Webster,
H. S. Scott, R. L. Boyd, L. Peltonen, and C. C. Goodnow. 2004. Gene dosage–
limiting role of Aire in thymic expression, clonal deletion, and organ-specific
autoimmunity. J. Exp. Med. 200: 1015–1026.
28. Nicolaides, N. C., S. J. Littman, P. Modrich, K. W. Kinzler, and B. Vogelstein.
1998. A naturally occurring hPMS2 mutation can confer a dominant negative
mutator phenotype. Mol. Cell. Biol. 18: 1635–1641.
29. Bourc’his, D., G. L. Xu, C. S. Lin, B. Bollman, and T. H. Bestor. 2001. Dnmt3L
and the establishment of maternal genomic imprints. Science 294: 2536–2539.
30. Bourc’his, D., and T. H. Bestor. 2004. Meiotic catastrophe and retrotransposon
reactivation in male germ cells lacking Dnmt3L. Nature 431: 96–99.
31. Shovlin, T. C., D. Bourc’his, S. La Salle, A. O’Doherty, J. M. Trasler,
T. H. Bestor, and C. P. Walsh. 2007. Sex-specific promoters regulate Dnmt3L
expression in mouse germ cells. Hum. Reprod. 22: 457–467.
1343The Journal of Immunology