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 autoimmune regulator (Aire3 protein) is thought to be a transcription factor (1) and, perhaps, an E3 ubiquitin ligase (2) that, when mutated, causes autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED) in humans (3, 4, 5, 6). In mice, Aire deficiency recapitulates many characteristics of APECED (7, 8, 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 promiscuous 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 infection, cell stress, or an “error catastrophe,” i.e., an accumulation of 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 investigated Aire expression in the testis, another tissue with promiscuous gene expression, and Aire’s role in the scheduled apoptosis in germ cells.

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, 13, 14), with a peak after 3 wk. Spermatocytes (12, 13) are thought to be affected most, with spermatogonia (14) also being 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 expression 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, sporadic 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 necessary for the maintenance of a critical ratio of cells of some germ cell stages to Sertoli cells (12, 13, 16). In this hypothesis, a presumed 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 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 deficiency 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.

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 experiments with Rag mice, HD Aire+/+ Rag−/− were compared with HD Aire−/− Rag−/−. The Morphomouse (21) was on an undefined background; 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.

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 according to standard procedures. Embedded testes were sectioned (4 μm) 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 Microscopy Sciences) using dry ice covered with 2-methyl butane. Frozen sections (4 μm) 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% H2O2 in methanol for 10 min at room temperature to inactivate endogenous peroxidase 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 substrate (Vector Laboratories) intensified with Ni2+ to obtain dark brown staining. Sections were counterstained with Nuclear Fast Red (Vector Laboratories) for 10 min, dehydrated in ethanol, cleared in xylene, mounted, and coverslipped.

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 incubated 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 immersion in 0.5% H2O2 in 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 chamber. 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′-diaminobenzidine (Vector Laboratories); counterstaining was done with Nuclear Fast Red (Vector Laboratories) for 5 min, and slides were mounted and coverslipped.

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 percentage of apoptotic cells. Germ cell stages of apoptotic cells were identified according to their morphology and position in the seminiferous tubule. 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.

Gene expression was measured using quantitative RT-PCR. RNA was extracted 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 performed using the ABI PRISM 7700 machine. Primers and probes were as follows:

Table I.

mRNA expression in mouse thymus and testis relative to β-actin mRNAa

MouseAIRE GenotypeAge (days)TissueAIRE (×10−4)FABP (×10−4)Hbb-y (×10−3)IRBP (×10−4)LTF (×10−4)NPY (×10−5)Insulin (×10−4)Ptdgs (×10−3)Spt1 (×10−5)Spt2 (×10−5)
−/− 23 Thymus Undb 1.7  Und 3.3  6.2 0.1 1.2  
−/− 23 Thymus Und 1.5  Und 0.8  15.0 0.1 1.0  
+/− 23 Thymus 102.9 2.9  Und 18.5  33.9 1.1 36.6  
+/− 23 Thymus 46.0 5.7  Und 11.5  34.3 0.6 55.8  
−/− 23 Testes Und Und  1.8 64.6  3.3 54.9 1.8   
−/− 23 Testes Und Und  1.5 17.0  3.2 40.8 0.7  
+/− 23 Testes 0.8 Und  2.4 0.0  3.5 44.4 0.8  
+/− 23 Testes 0.8 Und  1.7 4.2  3.2 21.7 0.9  
−/− 193 Testes Und Und 2.2 1.8 0.2 3.2 7.4 1446 7.3 9.0 
−/− 194 Testes Und Und 1.2 2.0 0.2 5.7 2.9 1212 3.8 5.8 
−/− 241 Testes Und Und 2.6 2.4 4.9 4.3 2.6 1618 5.2 6.7 
+/− 193 Testes 2.4 Und 1.4 1.3 19.8 4.0 2.0 1736 5.5 11.4 
+/− 194 Testes 3.0 Und 1.3 1.7 0.5 2.0 2.5 1538 4.3 16.1 
10 +/− 241 Testes 1.3 Und 0.9 1.9 0.2 0.7 3.1 840 5.2 11.0 
MouseAIRE GenotypeAge (days)TissueAIRE (×10−4)FABP (×10−4)Hbb-y (×10−3)IRBP (×10−4)LTF (×10−4)NPY (×10−5)Insulin (×10−4)Ptdgs (×10−3)Spt1 (×10−5)Spt2 (×10−5)
−/− 23 Thymus Undb 1.7  Und 3.3  6.2 0.1 1.2  
−/− 23 Thymus Und 1.5  Und 0.8  15.0 0.1 1.0  
+/− 23 Thymus 102.9 2.9  Und 18.5  33.9 1.1 36.6  
+/− 23 Thymus 46.0 5.7  Und 11.5  34.3 0.6 55.8  
−/− 23 Testes Und Und  1.8 64.6  3.3 54.9 1.8   
−/− 23 Testes Und Und  1.5 17.0  3.2 40.8 0.7  
+/− 23 Testes 0.8 Und  2.4 0.0  3.5 44.4 0.8  
+/− 23 Testes 0.8 Und  1.7 4.2  3.2 21.7 0.9  
−/− 193 Testes Und Und 2.2 1.8 0.2 3.2 7.4 1446 7.3 9.0 
−/− 194 Testes Und Und 1.2 2.0 0.2 5.7 2.9 1212 3.8 5.8 
−/− 241 Testes Und Und 2.6 2.4 4.9 4.3 2.6 1618 5.2 6.7 
+/− 193 Testes 2.4 Und 1.4 1.3 19.8 4.0 2.0 1736 5.5 11.4 
+/− 194 Testes 3.0 Und 1.3 1.7 0.5 2.0 2.5 1538 4.3 16.1 
10 +/− 241 Testes 1.3 Und 0.9 1.9 0.2 0.7 3.1 840 5.2 11.0 
a

Mice of the same age were siblings.

b

Und, Undetected expression.

AIRE: 5′-GCCAAGGGAGCCCAGG-3′, 5′-GGAGGAACCCCACACTGCT-3′, 5′-(6-FAM)-CACTATACCTGGTAGAGATGAGCAGAAAGTGGG-(TAMRA)-3′, Hbb-y: 5′-GCTAGTCACTTCGGCAATGAATT-3′, 5′-CCCCAGCCACCAGCTTC-3′, 5′-(6-FAM)- AGCTGAGATGCAGGCTGCCTGGC-(TAMRA)-3′; NPY: 5′-CAGAAAACGCCCCCAGAAC-3′, 5′-CGGGAGAACAAGTTTCATTTCC-3′, 5′-(6-FAM)-AGGCTTGAAGACCCTTCCATGTGGTGA-(TAMRA)-3′; Ptdgs: 5′-CCTGCCCCAACCGGAT-3′, 5′-GTGACCAGCCCTCTGACTGAC-3′, 5′-(6-FAM)-AGTGCATTCAAGAGTAAACGCAGGTGAGAG-(TAMRA)-3′; Spt1: 5′-GCTTGGTGTTTCCACTATCCTAGTCT-3′, 5′-AATCAGCAGTTCCAGAAGTTTCAGT-3′, 5′-(6-FAM)-TTGCCAGGACCCGGAGACAAACA-(TAMRA)-3′; Spt2: 5′-CACCATGAAGTTCCTGGCACT-3′, 5′-TCTCCGGGTCCTGGCAA-3′, 5′-(6-FAM)-CTTGTGTTGCTTGGTGTTTCCACTATCCTAGTC-(TAMRA)-3′; IRBP: 5′-AATGACTCGGTCAGCGAACTTT-3′, 5′-CTGTCACACCACTGGTCAGGAT-3′, 5′-(6-FAM)-ACAGGTGAACGATATGGCTCCAAG AAG-(TAMRA)-3′; FABP: 5′-CGTGTAGACAATGGAAAGGAGCT-3′, 5′-AAGAATCGCTTGGCCTCAACT-3′, 5′-(6-FAM)-TCATTACCAGAAACCTCTCGGACAGCA-(TAMRA)-3′; proinsulin: 5′-ATCTACAATGCCACGCTTCTG-3′, 5′-GACCCACAAGTGGCACAA-3′, 5′-(6-FAM)-GCCCGGGAGCAGGTGACCTT-(TAMRA)-3′; β-actin: 5′-AGGTCATCACTATTGGCAACGA-3′, 5′-CACTTCATGATGGAATTGAATGTAGTT-3′, 5′-(6-FAM)-TG CCACAGGATTCCATACCCAAGAAGG-(TAMRA)-3′.

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, 25, 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 differentiated by position within the tubule and their morphology (Fig. 1). In thymus, only medullary cells stained (Fig. 1,C), and there was no staining with irrelevant Ab (isotype control) (Fig. 1,A) or in tissue from Aire-deficient mice (Fig. 1,B). In testis, spermatogonia and spermatocytes stained, but not spermatids or Sertoli cells (Fig. 1,F), and no staining was observed in Aire-deficient tissue (Fig. 1,E) or with irrelevant Ab (isotype control) (Fig. 1 D). 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.

FIGURE 1.

Aire protein is expressed in the germ cells of the testis. Immunohistochemical stains on fixed frozen sections using mAb to mouse Aire protein. AC, Medullary region of thymus; DF, 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.

FIGURE 1.

Aire protein is expressed in the germ cells of the testis. Immunohistochemical stains on fixed frozen sections using mAb to mouse Aire protein. AC, Medullary region of thymus; DF, 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.

Close modal

In testis tissue from 3-wk-old Aire-sufficient littermates of LA Aire−/− and HD Aire−/− mice, ∼0.4% (0.43 and 0.41%, respectively) of cells were Aire positive, and in 3-mo-old mice, 0.15% of cells were Aire positive (Fig. 1,G). 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 β-actin expression was between 0.8 × 10−4 (3-wk-old) and 2.2 × 10−4 (3-mo-old) (Fig. 1 H). 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-β-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 underestimate, because we did not consider the presence of non-germ cells (e.g., Sertoli cells) in this analysis. We note, however, that there is no obvious defined differentiation state in which all the germ cells express Aire, implying sporadic expression. 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 developmental program. Along this line, Aire mRNA might simply be a passenger; in which case its translated product would be without function.

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 expression—“no effect” would read as 1.0. In the thymus (Fig. 2,A), 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. 2,B), the four genes were not influenced by Aire. In adult testis (Fig. 2 C), 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 subset 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 insulin 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).

FIGURE 2.

Dependence of gene expression on AIRE. A, 23-day-old thymus; B, 23-day-old testis; C, adult testis. AIRE dependence is represented as the quotient of gene expression levels in AIRE heterozygous and deficient mice. FABP, fatty acid binding protein 2 intestinal; Proins, proinsulin; Ptdgs, prostaglandin-D2 synthase; Spt1, salivary protein 1; IRBP retinol-binding protein 3 interstitial; Hbb-y, hemoglobin Y β-like embryonic chain; NPY, neuropeptide Y; Spt2, salivary protein 2 precursor. FABP was undetected in the 23-day testes, and IRBP was undetected in the thymus. The tests on their expression were performed but there was no interpretable signal in these after 40 cycles.

FIGURE 2.

Dependence of gene expression on AIRE. A, 23-day-old thymus; B, 23-day-old testis; C, adult testis. AIRE dependence is represented as the quotient of gene expression levels in AIRE heterozygous and deficient mice. FABP, fatty acid binding protein 2 intestinal; Proins, proinsulin; Ptdgs, prostaglandin-D2 synthase; Spt1, salivary protein 1; IRBP retinol-binding protein 3 interstitial; Hbb-y, hemoglobin Y β-like embryonic chain; NPY, neuropeptide Y; Spt2, salivary protein 2 precursor. FABP was undetected in the 23-day testes, and IRBP was undetected in the thymus. The tests on their expression were performed but there was no interpretable signal in these after 40 cycles.

Close modal

One of the dramatic events in spermatogenesis is scheduled apoptosis. Vaguely reminiscent of counterselection of T cells in the thymus, fewer than 20% of germ cells survive in this early apoptotic 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% apoptotic 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 independent of the adaptive immune system. The reduction was more pronounced when we differentiated between cell types: the spermatocytes II were affected the most, with only half as many apoptotic cells in LA (p = 0.08) and HD (p = 0.0003) Aire-deficient mice, compared with Aire-sufficient (Aire+/−) (Fig. 4).

FIGURE 3.

Percentage of apoptotic (TUNEL positive) germ cells. Bars represent the arithmetic mean with SD. Each data set is from 5 mice (except 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 patients with APECED, who often suffer from sterility associated with testicular atrophy (6 ), Aire-deficient mice have reduced fertility, with no anatomic abnormalities and no lymphocyte infiltration apparent in the testes (79 ). Nevertheless, the HD mice on a Rag-1-deficient background bred well.

FIGURE 3.

Percentage of apoptotic (TUNEL positive) germ cells. Bars represent the arithmetic mean with SD. Each data set is from 5 mice (except 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 patients with APECED, who often suffer from sterility associated with testicular atrophy (6 ), Aire-deficient mice have reduced fertility, with no anatomic abnormalities and no lymphocyte infiltration apparent in the testes (79 ). Nevertheless, the HD mice on a Rag-1-deficient background bred well.

Close modal
FIGURE 4.

Percentage of apoptotic germ cells, stratified according to 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.

FIGURE 4.

Percentage of apoptotic germ cells, stratified according to 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.

Close modal

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 apoptotic 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 germ cell stages seemed to be affected, from spermatogonia through spermatocytes to later stages (Fig. 4).

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 maintenance 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 hypothesis, 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 introduced into the cells of bacteria, yeasts, plants and mammalian cells, it increases the rate of genome-wide mutagenesis (21). Because 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 response to DNA damage (19), we thought that transgenic morphogene 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 abnormalities 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 morphogene reduces apoptosis and thus can interfere with the ratio of germ cells to Sertoli cells. Because this does not affect adult spermatogenesis 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 scheduled 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 apoptosis 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 necessity 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 rearrangements 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.

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 protein is unclear. In the mouse, tightly linked to and partially overlapping with the Aire locus is the Dnmt3L gene, transcribed in the opposite direction. Dnmt3L is a member of the DNA methyltransferase 3 family that lacks enzymatic activity but is required for de novo methylation of imprinted genes in oocytes (29) and for transposon 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 spermatocytes 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 recombination (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 counterselection 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 mutational load.

We thank Diane Mathis and Leena Peltonen for Aire-deficient mice; Nick Nicolaides for the morphogene mouse; Pärt Peterson for Abs; Nils Lonberg, Mindy Brooks, Nick Salmon and Renee Reijo Pera for discussions and support; and Margaret Mayes for technical help. This work was performed by C.S. as part of the requirement for the Ph.D. degree at the Ludwig-Maximilians-University Munich (supervisor, Dr. Elisabeth Weiss).

The authors have no financial conflict of interest.

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.

1

Supported by National Institutes of Health Grants R01 AI041570 and a BioSTAR grant funded by the University of California and Medarex (to M.W.).

3

Abbreviations used in this paper: Aire, autoimmune regulator; APECED, polyendocrinopathy-candidiasis-ectodermal dystrophy.

1
Kumar, P. G., M. Laloraya, C. Y. Wang, Q. G. Ruan, A. Davoodi-Semiromi, K. J. Kao, 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 APECED Consortium.
1997
. An autoimmune disease, 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, 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, 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, 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, D. Mathis.
2002
. Projection of an immunological self shadow within the thymus by the aire protein.
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, 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, 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, P. Vassalli.
1997
. An early and massive wave of germinal cell apoptosis is required for the development of functional spermatogenesis.
EMBO J.
16
:
2262
-2270.
13
Mori, C., N. Nakamura, D. J. Dix, M. Fujioka, S. Nakagawa, K. Shiota, 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
:
125
-136.
14
Wang, R. A., P. K. Nakane, 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, Y. Matsui.
1996
. Inhibition of testicular germ cell apoptosis and differentiation in mice misexpressing Bcl-2 in spermatogonia.
Development
122
:
1703
-1709.
16
Print, C. G., 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, 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
:
2420
-2425.
18
Chen, J., I. Sadowski.
2005
. Identification of the mismatch repair genes PMS2 and MLH1 as p53 target genes by using serial analysis of binding elements.
Proc. Natl. Acad. Sci. USA
102
:
4813
-4818.
19
Zeng, M., L. Narayanan, X. S. Xu, T. A. Prolla, R. M. Liskay, P. M. Glazer.
2000
. Ionizing radiation-induced apoptosis via separate Pms2- and p53-dependent 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, L. Grasso.
2005
. Morphogenics as a tool for target discovery and drug development.
Ann. NY Acad. Sci.
1059
:
86
-96.
22
Halonen, M., M. Pelto-Huikko, P. Eskelin, L. Peltonen, I. Ulmanen, 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, M. Harada.
2002
. Expression of AIRE gene in peripheral monocyte/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 medulla.
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, O. Tureci.
2006
. Expression profiling of autoimmune regulator AIRE mRNA in a comprehensive 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, 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, 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, T. H. Bestor.
2001
. Dnmt3L and the establishment of maternal genomic imprints.
Science
294
:
2536
-2539.
30
Bourc’his, D., 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, C. P. Walsh.
2007
. Sex-specific promoters regulate Dnmt3L expression in mouse germ cells.
Hum. Reprod.
22
:
457
-467.