- Split View
-
Views
-
Cite
Cite
Frédéric Flamant, Anne-Lise Poguet, Michelina Plateroti, Olivier Chassande, Karine Gauthier, Nathalie Streichenberger, Ahmed Mansouri, Jacques Samarut, Congenital Hypothyroid Pax8−/− Mutant Mice Can Be Rescued by Inactivating the TRα Gene, Molecular Endocrinology, Volume 16, Issue 1, 1 January 2002, Pages 24–32, https://doi.org/10.1210/mend.16.1.0766
- Share Icon Share
Abstract
Mice devoid of all TRs are viable, whereas Pax8−/− mice, which lack the follicular cells producing T4 and T3 in the thyroid gland, die during the first weeks of postnatal life. A precise comparison between the two types of mutants reveals that their phenotypes are similar, but the defects in spleen, bone, and small intestine are more pronounced in Pax8−/− mice. This is interpreted as the result of a negative effect of the unliganded TR on thyroid hormone target genes expression in the Pax8−/− mutants. Pax8/TRα compound mutants can survive to adulthood, and the expression of target genes is partially restored. This demonstrates the importance of TRα aporeceptor activity in several aspects of postnatal development.
THE TR GENES TRα and TRβ [NR1A1 and NR1A2 (Nuclear Receptors Nomenclature Committee, 1999)] belong to the superfamily of nuclear hormone receptor genes and encode several isoforms. Among these, TRα1, TRβ1, TRβ2, and TRβ3 (1) bind T3 and activate transcription in the nucleus. Like RAR, PPAR, VDR, farnesoid-X-activated receptor, liver X receptor, and Nurr1, the TRs form heterodimers with RXR and bind to DNA response elements located in transcription promoters (TRE). T3 directly activates gene transcription by binding to the C-terminal domain of TR, inducing a conformation change and the recruitment of transcription coactivators (2, 3). For a number of genes, T3 does not induce activation but, rather, induces transcriptional repression (4). Although the underlying mechanism is still unclear, it is usually assumed that it results from the negative action of liganded TR bound to specific TRE (5, 6).
The fact that DNA binding of TR is not hormone dependent also raises the possibility for a biological activity of the unliganded receptors, or aporeceptors (7). In vitro experiments show that unliganded TR/RXR heterodimers bound to DNA recruit transcription corepressors with histone deacetylase activity and exert a negative effect on TRE (8). However, the physiological relevance of this phenomenon has not been clearly evaluated. The recent availability of several mutant mice provides new possibilities to address this question in living animals.
Pax8 is a gene encoding a paired-box-containing protein with a highly restricted expression pattern, which is necessary for the differentiation of thyroid cells (9). The only known primary defect in Pax8−/− knockout mice is the absence of thyroid follicular cells (10). This defect results in the almost complete absence of T4 in postnatal life, when autonomous hormone production normally replaces maternal supply. Pax8−/− mice die around weaning time but can be rescued by T4 treatment. Several TR-knockout mice have also been produced (11, 12). Surprisingly, all single and compound TR-knockout mutants are viable, with one exception (13, 14), which is discussed elsewhere (15, 16). Thus, there is now ample evidence that mice devoid of all TR are viable (17) although very poorly fertile. Therefore, it seems that the absence of TR is less deleterious than the deficiency in T4 and T3 observed in Pax8−/− mice.
We combined the TRα0 mutation, which is viable and deletes all the TRα isoforms (15), and the lethal Pax8 mutation and found that animals homozygous for both knockout mutations can survive to adulthood. The lethality observed in Pax8−/− mice is thus a consequence of the TRα1 aporeceptor activity.
RESULTS
Pax8−/− Mice Have Similar But More Severe Defects Than TRα0/0β−/− Mice
The postnatal consequences of the absence of T4 and T3 synthesis in Pax8−/−-knockout mice have not been studied before. However, the fact that these mice do not survive beyond weaning time indicates that they are different from mice lacking TR (15, 17). We thus decided to perform a systematic comparison between the Pax8−/− mice and TRα0/0β−/− mice devoid of all TR isoforms, obtained by intercrossing TRα+/0TRβ−/− animals (15).
Statistical analysis reveals that Pax8−/− mice are present in Mandelian ratio at birth (9 of 35 newborns) but unlike TRα0/0β−/− mice, they are under-represented in offspring 2 wk after birth (Table 1). T4 is almost undetectable in Pax8−/− blood at this age, confirming that maternal supply is minimal after birth (data not shown). Early growth appeared quite variable but clearly slower in Pax8−/− mice than in TRα0/0β−/− mice. We did not observe the spontaneous survival of Pax8−/− mice beyond 30 d, but we were able to rescue these mutants by injecting T4 daily between d 15 and 21, as reported previously (10). This confirms that the lethality of Pax8−/− mice before or during weaning time is a direct consequence of hypothyroidism.
. | Pax8−/− . | Pax8−/− (+T4/T3)a . | TRα0/0β−/− . | Pax8−/−TRα0/0 . | Pax8−/−TRβ−/− . |
---|---|---|---|---|---|
Observed ratio (d 10–15) | 35 /218b | 25 /110 | 13 /132b | 39 /244b | |
Average weight (d 15) | 4.1± 0.9 | 5.2± 0.8 | 4.5± 1.2 | 3.8± 0.5 | |
Survival beyond d 21 | 0 /25 | 6 /10 | 15 /15 | 6 /8 | 0 /18 |
. | Pax8−/− . | Pax8−/− (+T4/T3)a . | TRα0/0β−/− . | Pax8−/−TRα0/0 . | Pax8−/−TRβ−/− . |
---|---|---|---|---|---|
Observed ratio (d 10–15) | 35 /218b | 25 /110 | 13 /132b | 39 /244b | |
Average weight (d 15) | 4.1± 0.9 | 5.2± 0.8 | 4.5± 1.2 | 3.8± 0.5 | |
Survival beyond d 21 | 0 /25 | 6 /10 | 15 /15 | 6 /8 | 0 /18 |
Seven days treatment starting at d 15.
Significantly inferior to the expected 25% ratio in heterozygous crosses (χ2 test, P = 0.05).
. | Pax8−/− . | Pax8−/− (+T4/T3)a . | TRα0/0β−/− . | Pax8−/−TRα0/0 . | Pax8−/−TRβ−/− . |
---|---|---|---|---|---|
Observed ratio (d 10–15) | 35 /218b | 25 /110 | 13 /132b | 39 /244b | |
Average weight (d 15) | 4.1± 0.9 | 5.2± 0.8 | 4.5± 1.2 | 3.8± 0.5 | |
Survival beyond d 21 | 0 /25 | 6 /10 | 15 /15 | 6 /8 | 0 /18 |
. | Pax8−/− . | Pax8−/− (+T4/T3)a . | TRα0/0β−/− . | Pax8−/−TRα0/0 . | Pax8−/−TRβ−/− . |
---|---|---|---|---|---|
Observed ratio (d 10–15) | 35 /218b | 25 /110 | 13 /132b | 39 /244b | |
Average weight (d 15) | 4.1± 0.9 | 5.2± 0.8 | 4.5± 1.2 | 3.8± 0.5 | |
Survival beyond d 21 | 0 /25 | 6 /10 | 15 /15 | 6 /8 | 0 /18 |
Seven days treatment starting at d 15.
Significantly inferior to the expected 25% ratio in heterozygous crosses (χ2 test, P = 0.05).
TSH activates T4 secretion and T3 synthesis, which exerts in turn a negative feedback regulation on pituitary TSH secretion. The number of TSH-secreting cells in pituitary is highly elevated in both TRα0/0β−/− and Pax8−/− mice (Fig. 1A), and the circulating level of TSH is always superior to 40 mU/liter (data not shown). TSHα gene expression was also found to be similar in these two genotypes by quantitative RT-PCR analysis (data not shown). Thus, although T4 and T3 levels are very low in Pax8−/− mice and very high in TRα0/0β−/− mice, the two situations are functionally equivalent.
Anatomical and histological analyses were thus performed 14–16 d after birth, with a particular attention for the organs that have been previously shown to be affected in various TR-knockout mice. Spleen hematopoiesis is impaired in TRα-knockout mice, mainly reflecting a defect in B cell development (18). TRα0/0β−/− animals display a visible splenic hypotrophy. This feature is much more pronounced in Pax8−/− mice (Fig. 1B), and histological examination confirms a dramatic reduction in the size and number of follicles (data not shown). Bone morphology was studied by x-ray, whole mount alizarine staining of skeleton, and on histological sections (Fig. 1, C and D). Both types of mutant display the defects reported previously for 3- to 6-wk-old TR-knockout mice (13, 17), i.e. retarded ossification and mineralization of long bones, which is clearly visible in epiphysis. These defects are likely to result from a defect in growth plate cell differentiation (19). Again, this phenotype was always more pronounced in Pax8−/− mice than in TRα0/0β−/− mice.
Small intestine epithelium is another important target for T3 during the profound remodeling occurring at weaning time (20). Epithelial cell proliferation and differentiation in the distal small intestine, assessed by immunostaining of the Ki67 proliferation marker and by measuring lactase activity, respectively, were more severely reduced in Pax8−/− than in TRα0/0β−/− mutants (Fig. 2 and Table 2). Cells were also much more vacuolized in Pax8−/− epithelium. When T4 and T3 were injected to Pax8−/− mice, a recovery of crypt cell proliferation was observed after 48 h. Morphological examination showed that the highly vacuolized epithelial cells, typical of immature stages (21), were replaced by apparently normal cells in the proximal portion of the villi (Fig. 2, lower panel). The only modest increase in epithelial lactase activity probably reflects the fact that the complete renewal of intestinal epithelium would take more than 48 h of hormonal treatment.
. | Pax8−/− . | Pax8−/− (+T4/T3)a . | TRα0/0β−/− . | Pax8−/−TRα0/0 . | Pax8−/−TRβ−/− . |
---|---|---|---|---|---|
Distal small intestine villi length (% wild type) | 67 ± 9b | 77 ± 3b | 80 ± 8b | 93 ± 5c | 65 ± 4b |
Distal small intestine proliferation (% wild type) | 33 ± 10b | 69 ± 8b | 50 ± 11b | 83 ± 7bc | 42 ± 7b |
Distal small intestine lactase activity (% wild type) | 71 ± 7b | 75 ± 7b | 92 ± 3 | 93 ± 1c | 62 ± 2b |
Femur length (% wild type) | 65 ± 15 | 60 ± 13 | 93 ± 5 | 80 ± 12c | 63 ± 10 |
. | Pax8−/− . | Pax8−/− (+T4/T3)a . | TRα0/0β−/− . | Pax8−/−TRα0/0 . | Pax8−/−TRβ−/− . |
---|---|---|---|---|---|
Distal small intestine villi length (% wild type) | 67 ± 9b | 77 ± 3b | 80 ± 8b | 93 ± 5c | 65 ± 4b |
Distal small intestine proliferation (% wild type) | 33 ± 10b | 69 ± 8b | 50 ± 11b | 83 ± 7bc | 42 ± 7b |
Distal small intestine lactase activity (% wild type) | 71 ± 7b | 75 ± 7b | 92 ± 3 | 93 ± 1c | 62 ± 2b |
Femur length (% wild type) | 65 ± 15 | 60 ± 13 | 93 ± 5 | 80 ± 12c | 63 ± 10 |
Forty-eight hours treatment.
Significantly reduced compared to wild type (t test, P = 0.05).
Significantly different from Pax8−/−.
. | Pax8−/− . | Pax8−/− (+T4/T3)a . | TRα0/0β−/− . | Pax8−/−TRα0/0 . | Pax8−/−TRβ−/− . |
---|---|---|---|---|---|
Distal small intestine villi length (% wild type) | 67 ± 9b | 77 ± 3b | 80 ± 8b | 93 ± 5c | 65 ± 4b |
Distal small intestine proliferation (% wild type) | 33 ± 10b | 69 ± 8b | 50 ± 11b | 83 ± 7bc | 42 ± 7b |
Distal small intestine lactase activity (% wild type) | 71 ± 7b | 75 ± 7b | 92 ± 3 | 93 ± 1c | 62 ± 2b |
Femur length (% wild type) | 65 ± 15 | 60 ± 13 | 93 ± 5 | 80 ± 12c | 63 ± 10 |
. | Pax8−/− . | Pax8−/− (+T4/T3)a . | TRα0/0β−/− . | Pax8−/−TRα0/0 . | Pax8−/−TRβ−/− . |
---|---|---|---|---|---|
Distal small intestine villi length (% wild type) | 67 ± 9b | 77 ± 3b | 80 ± 8b | 93 ± 5c | 65 ± 4b |
Distal small intestine proliferation (% wild type) | 33 ± 10b | 69 ± 8b | 50 ± 11b | 83 ± 7bc | 42 ± 7b |
Distal small intestine lactase activity (% wild type) | 71 ± 7b | 75 ± 7b | 92 ± 3 | 93 ± 1c | 62 ± 2b |
Femur length (% wild type) | 65 ± 15 | 60 ± 13 | 93 ± 5 | 80 ± 12c | 63 ± 10 |
Forty-eight hours treatment.
Significantly reduced compared to wild type (t test, P = 0.05).
Significantly different from Pax8−/−.
The overall conclusion of the above observations is that Pax8−/− and TRα0/0β−/− mutant mice display similar phenotypic traits, but that these defects are always more pronounced in the Pax8−/− animals.
Survival Without Thyroid Gland and TR
At this point, three main hypotheses could account for our observations: 1) A hypothetical third TR gene would encode a receptor, which would transduce T3 signal in some TRα0/0β−/− tissues and introduce a partial functional compensation. It has been shown that knocking out both the TRα and TRβ genes was sufficient to abrogate any high-affinity T3 binding in adult liver (17). Furthermore, there is at this time no indication for the presence of an uncharacterized TR-related sequence in the GenBank database (Robinson-Richevi, M., and V. Laudet, unpublished observations). The existence of a third unknown TR gene, expressed in other tissues, is therefore highly unlikely. 2) Nongenomic effects of thyroid hormones might persist in TRα0/0β−/− but not in Pax8−/− animals to ensure survival. The physiological relevance of such effects are controversial and the underlying mechanism is poorly understood, but T4 is able to act directly on cell membranes and cytoplasmic components to activate an alternative signaling pathway (22–24). 3) TR aporeceptors present in the Pax8−/− mice may have a detrimental physiological activity. According to in vitro data, TR/RXR heterodimers and TR/TR homodimers (25) act as transcription repressors in the absence of T3. Thus, TR target genes’ expression would be more sensitive by the absence of T3 than by the absence of receptors.
This third hypothesis leads to a testable prediction: compound mutant mice, which would lack both the thyroid hormones and the receptors, should be viable with a phenotype similar to the one observed in TRα0/0β−/− mice. To test this prediction, we crossed the TR and Pax8 mutants to generate Pax8−/−TRα0/0 and Pax8−/−TRβ−/− mice (Table 1).
The Pax8−/−TRβ−/− combination was lethal before weaning (Table 1 and Fig. 3). Anatomical and histological examinations failed to reveal a significant difference between these mice and Pax8−/− mice, except in one case (not shown), in which femural ossification was slightly more advanced in the compound mutant. In particular, the distal small intestine, bones, and spleen were always deeply affected (Table 2 and Fig. 4). This feature is not surprising, considering that previous knockout analysis failed to identify any function of TRβ in these organs (14, 26). Histological analysis of pituitary revealed the presence of a high number of TSH-secreting cells, similar to the one already observed for in Pax8−/− and TRα0/0β−/− (data not shown). This observation argues against any major influence of the TRβ aporeceptor on TSHα and TSHβ genes expression.
The offspring of Pax8+/−TRα+/0 and Pax8+/−TRα0/0 mice contained both Pax8−/−TRα+/0 and Pax8−/−TRα0/0 pups. The majority of Pax8−/−TRα0/0 mice survived to adulthood (Table 1), although growth was delayed compared with mice lacking only TR (Fig. 3). Anatomical and histological examinations confirmed a partial to complete phenotypic recovery in bone, spleen, and small intestine (Table 2 and Fig. 4). Surprisingly, some of the Pax8−/−TRα+/0 mice (three of eight) also spontaneously survived to adulthood, indicating that the absence of a single TRα gene copy limits the aporeceptor detrimental effect.
Attempts to obtain triple Pax8−/−TRα0/0TRβ−/− mutant mice were hampered by the very low fertility of double mutants. Only one Pax8−/−TRα0/0TRβ−/− male was obtained, which survived to adulthood and was not studied further.
From these observations, we can conclude that the lethality observed in Pax8−/− mutant mice is due to the negative effect of the TRα1 aporeceptor.
Direct in Vivo Evidence for TR Aporeceptors-Mediated Gene Regulation
T3 target genes in the small intestines, bone, and spleen are not known. Establishing a direct link between the presence of TR aporeceptors and the repression of T3 target genes’ expression in these tissues is therefore problematic. By contrast, although knockout of TR genes does not induce any obvious histological defect in liver and heart, their function is known to be regulated by T3. Several T3 target genes expressed in these two organs are now well characterized, and we decided to measure the expression of two of these. Liver cells express both TRα and TRβ genes, but TRβ mRNA is more abundant than TRα mRNA. The opposite situation is found in the heart. This is consistent with the results of knockout mice analysis, which revealed that TRβ has a predominant function in liver, whereas TRα is the main regulator of heart rate (27–30).
Among the best characterized T3 targets in liver is the gene encoding type I deiodinase gene (D1), which is very tightly regulated (31, 32) and which, in humans, contains several TRE in its promoter (33). As predicted, expression of D1 is greatly reduced in TRα0/0β−/− mice, undetectable in Pax8−/− mice, and highly induced after T4/T3 treatment (Fig. 5A). The persistence of D1 expression in TRα0/0β−/− livers is a good indication that unliganded TR can repress D1 transcription by binding to positive TRE in Pax8−/− livers. The fact that D1 expression is not detectable in either Pax8−/−TRβ−/− mice or Pax8−/−TRα0/0 mice is also an indication that both TRα and TRβ aporeceptors can contribute to this repression.
HCN2 is a gene encoding one of the pacemaker current components, which has been recently identified as a T3 target in cardiac cells, although the binding of TR to the promoter has not been demonstrated (34). HCN2 expression followed the expected expression pattern in heart with a mRNA level that is very low in Pax8−/−, higher in TRα0/0β−/− hearts, and inducible by T4/T3 treatment. HCN2 expression was clearly restored in Pax8−/−TRα0/0 mice but not in Pax8−/−TRβ−/− mice. Variations within mice sharing the same phenotype were very limited (Fig. 5C). These data are consistent with a repression of HCN2 transcription by the TRα1, but not the TRβ, aporeceptor.
DISCUSSION
The purpose of this work was to address the possible biological activity of unliganded TR in vivo. A key observation is that mice lacking a thyroid gland can survive only if they do not express the TRα gene. Thus, the TRα1 aporeceptor is able to compromise postnatal survival.
Congenital Hypothyroidism Is Lethal in Mice
Severe congenital hypothyroidism, as observed in Pax8−/− mice, is lethal at weaning time. T4 treatment during the second week of postnatal life is only sufficient to rescue Pax8−/− mice, showing that T4 and T3 are dispensable for survival after weaning. A similar lethality has been observed in other transgenic mice expressing a mutated cAMP response element binding protein gene in thyroid follicular cells in which T4 production is abrogated by a less direct mechanism (35). By contrast, drug treatment (36) and TSH receptor mutations (37, 38) induce a hypothyroid status that is not lethal. These discrepancies are likely to result from a moderate reduction of T3 levels in these last cases. The exact causes for the lethality linked to the Pax8-knockout remain hypothetical. In agreement with previous observations that showed that only TRα is important in this process (16), we found that the delay in small intestine development is much more pronounced in the lethal Pax8−/− and Pax8−/−TRβ−/− mutants than in all other genotypes. This delay is certainly sufficient to explain why Pax8−/− and Pax8−/−TRβ−/− mice do not survive beyond the weaning period, which normally occurs during the third postnatal week. These mutants are, however, already under-represented in offspring and underweight at d 10–15. T3 may therefore fulfill other unidentified vital functions during the first days of postnatal life in bone, spleen, heart, or other organs that were not studied here. Finally, because Pax8−/− embryos receive hormones through the placenta of their Pax8+/− mothers, we can not rule out another vital function for T3 during embryonic and fetal life.
TR Mutations Recapitulate Hypothyroidism
The resemblance between Pax8−/−- and TRα0/0β−/−-knockout mice suggests that the defects observed in bone, small intestine epithelium, and spleen directly reflect congenital hypothyroidism. This confirms that none of the phenotypic traits that we have observed are side effects of the Pax8 mutation per se. Meanwhile, it argues against a possible involvement of the nonreceptor isoforms encoded by the TR genes in the determination of these phenotypic traits (39, 40).
TRα1 and TRβ Aporeceptor Activity
The new observations presented here establish the first in vivo evidence that the repressor activity of the TRα1 aporeceptor is of physiological relevance. We were able to directly verify this assumption only for the HCN2 gene expression in heart. Unfortunately, the TRE responsible for the activation of this gene by T3 are presently unknown and other known T3 target genes usually do not display the same high induction rate. We were thus unable to address the possibility that T3 target genes may differ in their sensitivity to TRα1 aporeceptor. Based on histological and functional analysis, we can, however, predict that genes activated by T3 are actively repressed in Pax8−/− bone, spleen, and intestinal epithelium and that this repression is abrogated in Pax8−/−TRα0/0 mice. The mutant collection that we produced may then serve to identify some of these genes.
The large similarities observed between the phenotypes of Pax8−/− and Pax8−/−TRβ−/− mice does not provide any firm indication for TRβ aporeceptors activity. However, the persistence of D1 expression in TRα0/0β−/− but not Pax8−/− mice suggests that this gene is sensitive to both TRα1 and TRβ aporeceptors. D1 expression has recently been reported to be absent in mice devoid of TR, but this discrepancy might result from differences in the detection sensitivity achieved by Northern blotting (41). According to published data (29), other T3 target genes in the liver do not follow the same expression pattern, and T3 regulation in this organ seems to be very complex (4). The expression of D1 in TRα0/0β−/− liver could thus be interpreted in several different manners. According to recent data, TRβ aporeceptors exert a physiological function mainly in brain (42). This conclusion was reached after the introduction of a point mutation in the activation function-2 domain of TRβ performed to mimic the human syndrome of resistance to thyroid hormone in mice. Even mice heterozygous for this mutation display cerebellar and hippocampal abnormalities presumably absent in knockout mice and therefore attributed to TRβ aporeceptors activity. However, this interpretation remains questionable, as a slightly different mutation provides different results (43). Therefore, the possibility remains than TRα and TRβ aporeceptors activities are intrinsically different (44).
Possible Physiological Implication of Aporeceptor-Mediated Gene Regulation
The detrimental effects of TR aporeceptors on mouse postnatal development mimics human pathological situations such as thyroid disgenesis (45), agenesis (46), and perhaps resistance to thyroid hormone (47). A remaining issue is whether the TR aporeceptors also regulate transcription in a normal situation. There is no firm indication that T3 can be completely absent in any tissue expressing TRα or TRβ in healthy animals, but this question should be examined more precisely in the future. For example, T3 is not transported in brain but is produced locally by T4 deiodination. Deiodination in brain is performed by type 2 deiodinase, an enzyme with a restricted distribution (48) and under tight regulation (49, 50). T3 could therefore be absent in some brain areas during the fetal or postnatal life. T3 may also be unable to access to some embryonic tissues at a time when embryos already express TRα.
Do Other Aporeceptors Regulate Gene Transcription in Vivo?
The TR/T3 pathway offers a favorable opportunity to verify the physiological relevance of aporeceptors activity by genetic means. First, there is a well-defined source of ligand in the organism, and the Pax8 mutation provides an extremely precise way to perform a genetic ablation. Second, all the receptors isoforms are encoded by only two genes. It is likely, however, that aporeceptor function also exists for other related receptors. For example, it might explain why VDR knockout is not equivalent to vitamin D deficiency (51) or why retinaldehyde dehydrogenase type 2 knockout, which abrogates RAR ligand synthesis, entails early embryonic lethality (52), whereas RAR compound mutants usually survive beyond this stage (53). The recent knockout of nuclear corepressor 1 also suggests that many nuclear aporeceptors have important developmental functions (54), although this corepressor interacts at least in vitro with a broad range of transcription factors.
In conclusion, this work demonstrates that TRα1 aporeceptor have physiological effects in vivo and that T3 is required to relieve these effects during the first weeks of postnatal development.
MATERIALS AND METHODS
Mice
All pups were kept with their mother for at least 5 wk to ensure natural weaning. Genotypes were determined on toe lysates at d 10–15 by PCR DNA analysis. The TRα0 allele (15) is a complete deletion of all the known TRα isoforms. It differs from the TRα− allele used in previous studies (13) by the fact that it is viable and fertile. TRα0/0 animals were analyzed with a mixture of 4 oligonucleotides (5′-ATCGCCTTCTATCGCCTTCTTGACG, 5′-TTCAGGAGGATGATCTGGTCTTCGCAAG, 5′-GAGGAGGCGAAAGGAGGAG, and 5′-TGCCCTGGGCGTTAGTGCTG) and the following temperature schedule: 95 C for 5 min (94 C for 20 sec, 60 C for 20 sec, 72 C for 60 sec) 5 times; (94 C for 20 sec, 56 C for 20 sec, 72 C for 60 sec) 27 times to amplify both the wild type (115 bp) and the mutant (660 bp) allele. The PCR protocol to detect the TRβ mutation was reported previously (14). The Pax8-knockout mutation was described before (10). The genotype at this locus was determined using 3 oligonucleotides (5′-GGATGTGGAATGTGTGCGAGG, 5′-GCTAAGAGAAGGTGGATGAGAG, and 5′-GATGCTGCCAGTCTCGTAG) and the following temperature schedule: 94 C for 5 min (94 C for 15 sec, 60 C for 15 sec, 72 C for 30 sec, 10 times; 94 C for 15 sec, 57 C for 15 sec, 72 C for 30 sec, 25 times), 72 C for 5 min. This amplifies a 390-bp fragment for the wild-type allele and a 370-bp fragment for the knockout allele. All genetic background were initially a combination of C57/BL and 129Sv later backcrossed to 129Sv. Wild-type controls were littermates of Pax8 mutants. Experiments were all performed with young animals kept with their mother. When indicated, thyroid hormones [1 μg T3 and 10μ g T4 (Sigma, St. Louis, MO) in 100 μl PBS] were injected ip at d 13 and 14, and animals were sacrificed at d 15.
Histology and Enzymatic Activities
Standard histological techniques were used (55). Distal small intestine (56) and bone (13) were analyzed as described. Values of sd were calculated from at least 3 animals. For TSH immunocytochemistry, pituitary glands were fixed in Bouin-Hollande-Sublimate for 4 d and then embedded in parrafin. Serial sections were treated by the indirect immunoperoxidase method with streptavidin-biotin-complex (DAKO Corp. A/S, Copenhagen, Denmark) using a 1/20000 dilution of an anti-βrTSH antibody (a kind gift from Dr. A. F. Parlow, NIDDK) as described previously (57).
RNA Analysis
RNA were purified using column purification (QIAGEN, Valnecia, CA). Liver RNA was analyzed by Northern blotting using Hybond N+ membrane for transfer (Amersham Pharmacia Biotech, Piscataway, NJ) and an antisense RNA probe synthesized from a cloned D1 cDNA with 32P-labeled UTP (Amersham Pharmacia Biotech) for hybridization. HCN2 mRNA in heart was measured using a RNase protection assay (Ambion, Inc., Austin, TX) including a hypoxanthine phosphoribosyltransferase antisense probe for calibration. Quantitation was performed using a PhosphorImager and the ImageQuant software (Molecular Dynamics, Inc., Sunnyvale, CA).
Acknowledgments
This work was supported by the Max Planck Society, the Association pour la Recherche contre le Cancer, the Comité Départemental de la Loire de la Ligue Nationale contre le Cancer, and the Human Frontier scientific Program (RGO347/1999.M).
We thank Jean-Paul Roux for bone sections; Samuel Refetoff for helpful discussion; Bernd Gloss for the HNC2 probe gift; and Nadine Aguilera, Djamel Belgarbi, and Christelle Morin for animal breeding.
These authors contributed equally to this work.