|Home | About | Journals | Submit | Contact Us | Français|
Tuberous sclerosis complex (TSC) is a tumor suppressor gene syndrome in which hamartomas develop in multiple organ systems. Knockout and conditional alleles of Tsc1 and Tsc2 have been previously reported. Here, we describe the generation of a novel hypomorphic allele of Tsc2 (del3), in which exon 3, encoding 37 amino acids near the N terminus of tuberin, is deleted. Embryos homozygous for the del3 allele survive until E13.5, 2 days longer than Tsc2 null embryos. Embryos die from underdevelopment of the liver, deficient hematopoiesis, aberrant vascular development and hemorrhage. Mice that are heterozygous for the del3 allele have a markedly reduced kidney tumor burden in comparison with conventional Tsc2+/− mice. Murine embryo fibroblast (MEF) cultures that are homozygous for the del3 allele express mutant tuberin at low levels, and show enhanced activation of mTORC1, similar to Tsc2 null MEFs. Furthermore, the mutant cells show prominent reduction in the activation of AKT. Similar findings were made in the analysis of homozygous del3 embryo lysates. Tsc2-del3 demonstrates GTPase activating protein activity comparable to that of wild-type Tsc2 in a functional assay. These findings indicate that the del3 allele is a hypomorphic allele of Tsc2 with partial function due to reduced expression, and highlight the consistency of AKT downregulation when Tsc1/Tsc2 function is reduced. Tsc2-del3 mice also serve as a model for hypomorphic TSC2 missense mutations reported in TSC patients.
Tuberous sclerosis complex (TSC) is a human genetic disease characterized by the development of complex tumors termed hamartomas in multiple organ systems (1,2). During early childhood, the neurological manifestations of TSC are the most important clinical issue, and include seizures, developmental delay, autism and related phenotypes and mental retardation (3). However, most patients survive well past teenage years. In later life, renal angiomyolipomas and pulmonary lymphangioleiomyomatosis become major sources of morbidity and early mortality (1,4,5). Inactivating mutations in either of two genes, TSC1 and TSC2, cause this autosomal dominant tumor suppressor gene syndrome (6).
There is a wide spectrum of common mutations in the TSC1 and TSC2 genes, including large genomic deletions, nonsense mutations, splice site mutations, indel mutations and missense mutations (1,6). The majority of these mutations cause complete loss of function of one allele of TSC1 or TSC2, which when combined with a second hit event, typically a large genomic deletion at the other allele of the same locus, results in development of a TSC hamartoma (6). Complete loss of expression of either gene leads to the same biochemical and growth defect, as the two proteins form a heterodimeric complex. However, an increasing number of TSC2 missense mutations are known to result in the expression of a mutant TSC2 protein which is stably expressed, forms a complex with TSC1, and in some cases appears to have partial function (7,8). In addition, a few TSC2 missense mutations have been associated with a milder form of TSC in patients (9–11), suggesting the possibility that partial function hypomorphic alleles lead to a milder clinical form of the disease.
The TSC1/TSC2 protein complex functions as a critical integrator of growth signaling pathways within the cell that control the state of activation of an ancestrally conserved protein complex termed mTORC1 (12,13). The mTOR kinase present in activated mTORC1 phosphorylates its downstream targets, the S6 kinases and 4EB proteins, leading to cell size enlargement, ribosome biogenesis and increased protein synthesis. Loss of TSC1/TSC2 is also known to reduce the state of activation of AKT, at least partially through effects on the level of phosphorylation at the regulatory S473 site. This reduced phosphorylation due to loss of TSC1/TSC2 is thought to occur through effects on the level of expression of insulin receptor substrate (IRS), platelet-derived growth factor receptor (PDGFR) and activity of mTORC2 (14–16).
Multiple mouse models of TSC have been generated, including null and conditional alleles of Tsc1 and Tsc2 (17–24). Here we describe a novel conditional hypomorphic allele of Tsc2 in which exon 3 near the N terminus of the protein is conditionally deleted, resulting in reduced expression of a variant Tsc2 lacking the 37 amino acids encoded by exon 3.
To generate mice with a conditional allele of Tsc2, we employed a conventional gene targeting strategy (Fig. 1A). We chose to target exon 3 of Tsc2 which consists of 111 bp, encoding 37 amino acids near the N terminus of the Tsc2 protein (note that in some schemes this is counted as exon 4; however, extensive mutational analyses of TSC2 have always denoted this as exon 3, labeling a 5′ untranslated exon as 1a). Extensive sequence and computational analysis of the targeted region indicated that the inserted loxP sites did not affect splice site or exonic sequences. Chimeric founder mice were readily obtained, and bred with wild-type C57BL/6 mice to initiate a colony. Mice heterozygous for the conditional allele, hereafter termed c-del3, had no phenotype and were readily interbred to obtain homozygous Tsc2c-del3/c-del3 mice. Tsc2c-del3/c-del3 mice also had no phenotype, with survival out past 18 months, and normal breeding and behavioral characteristics (Fig. 1B).
A deleted allele with loss of exon 3, hereafter denoted the del3 allele, was generated from the conditional allele by recombination in vivo. A similar mixed strain colony of Tsc2del3/+ mice was also established.
No Tsc2del3/del3 pups were obtained at birth from matings of Tsc2del3/+ parents (Table 1). Analysis of timed matings demonstrated that Tsc2del3/del3 embryos appeared to be fully viable through E13.5, occurring in Mendelian ratios with little embryonic loss, but were not viable at E14.5 or later (Fig. 1C and Table 1). Note that the timing of embryonic death for Tsc2del3/del3 embryos was significantly later than that seen in Tsc2−/− embryos for which the majority of embryos were absent or resorbing by E11.5, and no viable embryos were seen at age E13.5 (17). Therefore, an analysis for the cause of death focused on E13.5 embryos. Tsc2del3/del3 E13.5 embryos were 0.5–1 mm smaller and developmentally retarded by 1–2 Theiler stages (25) in comparison with control embryos of either Tsc2+/+ or Tsc2del3/+ genotypes (Fig. 2A and D). However, placental vasculature and structure in Tsc2del3/del3 embryos appeared to be normal. In contrast, five out of five embryos that were examined in serial sections showed evidence of hemorrhage at multiple sites, including liver, brain and heart at E13.5 (Fig. 2B and C, Supplementary Material, Fig. S1A, C, D). Hematoxylin and eosin (H&E) stains as well as immunohistochemical stains for endomucin (26) demonstrated that blood vessels of the cutaneous tissues, brain and cardiac regions all appeared larger and to have aberrant branching morphology in Tsc2del3/del3 embryos in comparison with control embryos of either Tsc2+/+ or Tsc2del3/+ genotypes (Fig. 2D–F). Hemorrhage was seen in the same areas as the dilated blood vessels in these embryos, and appeared likely to be due to rupture from these aberrant structures (Supplementary Material, Fig. S1A).
Similar to findings made previously in both Tsc1 null and Tsc2 null embryos (17,20), the liver of Tsc2del3/del3 embryos was much smaller in comparison with control littermates (Fig. 2B). In addition, livers from Tsc2del3/del3 embryos had an abnormal architecture, with regions of necrosis and apoptosis near hematopoietic islands (Fig. 2C, Supplementary Material, Fig. S1C). Apoptosis was not widespread in these embryos overall, being seen only in the liver and near regions of hemorrhage. On the other hand, increased numbers of immature blood cells were seen in vascular channels in the Tsc2del3/del3 embryos, in comparison with control littermates (Supplementary Material, Fig. S1B). These observations in aggregate suggest that underdevelopment of the liver accompanied by poor/deficient hematopoiesis occurs in the Tsc2del3/del3 embryo, which in turn may lead to accelerated/aberrant vascular development, hemorrhage and death. In addition, they provide convincing evidence that the del3 allele has a phenotype intermediate between that of a complete null allele and a wild-type allele; it is a hypomorphic allele of Tsc2.
In contrast, nervous system development appeared normal in Tsc2del3/del3 embryos at E13.5. Nestin, NeuN, MBP expression and cell proliferation (as assessed by Ki-67 staining) were all similar in Tsc2del3/del3 and control embryos (Supplementary Material, Fig. S2).
To assess the in vivo function of the del3 allele in tumor development, we generated cohorts of Tsc2del3/+ mice. Since strain effects on tumor development in Tsc2+/− mice are well known (17,21 and unpublished data), we performed an intercross between Tsc2del3/+ and Tsc2+/− mice, and evaluated littermate cohorts of mice with the two genotypes for tumor development. Both gross and microscopic tumor scores were markedly reduced in the Tsc2del3/+ mice compared with the Tsc2+/− mice (gross tumor scores were 14.3-fold less, on average, P = 0.0004; microscopic tumor scores were 12.4-fold less, on average, P < 0.0001; Fig. 3A and B). The percent cellularity of the lesions (see Materials and Methods for details) was also reduced in the Tsc2del3/+ mice compared with the Tsc2+/− mice (51% versus 18% cellular on average, P < 0.0001, Fig. 3C). Qualitatively, the tumors from the Tsc2del3/+ mice were obviously smaller in size and growth appearance, and at age 18 months became comparable to those seen in the Tsc2+/− mice at age 12 months (Supplementary Material, Fig. S3).
We examined the Tsc2del3/+ and Tsc2+/− mouse kidney tumors to attempt to determine the reason for their difference in growth and overall size. pS6(S235/236) staining appeared somewhat higher in the cystadenomas of the Tsc2+/− mice compared with the Tsc2del3/+ mice (Supplementary Material, Fig. S4). Both Ki-67 and Cyclin D1 expression were also more common in tumor cells, on average, in the cystadenomas of Tsc2+/− mice compared with the Tsc2del3/+ mice (P = 0.05 and P = 0.04, respectively, Fig. 3D and E; Supplementary Material, Fig. S4). In contrast, pAKT(S473) expression was similarly low in lesions from mice of each genotype (Supplementary Material, Fig. S4).
We generated murine embryo fibroblast (MEF) cultures to study the signaling and growth characteristics of cells containing the del3 allele. MEF cultures of all three genotypes (Tsc2+/+, Tsc2del3/+, Tsc2del3/del3) were established, and immortalized at low frequency following a 3T3 culture protocol (15). Three immortalized Tsc2del3/del3 MEF lines were obtained.
Tsc2del3/del3 MEF cultures showed variably reduced (25–50% of normal) expression of Tsc2 in comparison with control wild-type MEFs derived simultaneously from littermate embryos, but had higher levels of Tsc2 than Tsc2 null MEF lines previously generated (15) (Fig. 4A). Quantitative RT–PCR analyses demonstrated that Tsc2 mRNA levels in Tsc2del3/del3 MEFs were 70–80% of levels seen in control MEF lines. Sequence analysis of Tsc2 cDNA from Tsc2del3/del3 MEFs demonstrated that there was a clean deletion of exon 3, as expected from the targeting construct.
Since the size of the predicted protein product was 37 amino acids or 4 kDa less than that of wild-type Tsc2 (200 kDa), we did not expect to see a migration difference on SDS–PAGE gels. However, there was a hint of such a difference on 4–12% acrylamide gels and on 6% acrylamide gels, there was a clear migration difference between the wild-type Tsc2 protein and that derived from the del3 allele (Tsc2-del3) (Fig. 4B). Immunoprecipitation studies demonstrated that the mutant Tsc2-del3 protein formed a stable complex with Tsc1, similar to wild-type Tsc2 (Fig. 4C). However, again, levels of Tsc2-del3 were variably reduced in comparison with the control lines.
To examine the cause of the reduced levels of Tsc2-del3, the MEF cell lines were treated with the protein synthesis inhibitor cycloheximide. Following cycloheximide treatment for 2 h, Tsc2-del3 protein levels (assessed in three Tsc2del3/del3 lines) declined by 35% in comparison with a negligible decline in Tsc2 protein levels in similarly treated wild-type fibroblasts (data not shown). To examine this in greater detail, we expressed Tsc1 and Tsc2 in HEK293 cells. We found that using the same vectors, the dosage of Tsc2 plasmid required to get expression equivalent to that of Tsc2-del3 was 8-fold lower (data not shown). However, when expressed at equivalent levels, Tsc2-del3 protein in HEK293 cells had similar stability (perhaps modestly reduced) to that of wild-type Tsc2 following cycloheximide treatment (Fig. 4D). These data suggest that the low expression of Tsc2-del3 is a result of reduced mRNA levels and reduced translation of its mRNA.
We explored the signaling characteristics of the Tsc2del3/del3 MEFs in comparison with control wild-type cells and Tsc2−/− MEFs (15). Phosphorylation of S6Kinase at the T389 [mTORC1 site (13)] was increased in the Tsc2del3/del3 MEFs in comparison with control lines during serum starvation, but was less than what was seen in Tsc2−/− MEFs (Fig. 5A). Levels of pS6(S235/236) and pS6(S240/244), downstream of S6kinase, were also increased in the Tsc2del3/del3 MEFs in comparison with controls, but were less than in the Tsc2−/− MEFs (Fig. 5A). Changes in S6 phosphorylation were quantified by gel densitometry, and the pS6:S6 level was determined (Fig. 5E). By this measure, the molar level of pS6 phosphorylation was comparable between the Tsc2del3/del3 and Tsc2−/− cell lines, reflecting higher total levels of S6 in the Tsc2−/− cell lines. Serum-stimulated phosphorylation of AKT at S473 was markedly reduced in the Tsc2del3/del3 MEFs in comparison with controls, but was somewhat higher than that seen in Tsc2−/− MEFs (Fig. 5A). Thus, in both mTORC1 activation and phosphorylation of Akt in response to serum, the Tsc2del3/del3 MEFs had an intermediate phenotype between wild-type and Tsc2−/− MEFs, but were much closer to the Tsc2−/− MEFs.
We then asked whether the intrinsic GTPase activating protein (GAP) activity of Tsc2-del3 protein was reduced in comparison with that of wild-type Tsc2 using a HEK293 assay, as used previously by others (9,16). When TSC1, S6K1 and Tsc2 or Tsc2-del3 are co-expressed in HEK293T cells, pS6K1 levels serve as a sensitive indicator of Tsc2 GAP activity. As above, we used differing doses of the Tsc2 plasmids to achieve similar levels of expression. We found that Tsc2-del3 had GAP activity that was similar to that of full-length Tsc2 (Fig. 5B), suggesting that the increased mTORC1 activity in the Tsc2del3/del3 cells was due to a lower level of expression of the Tsc2-del3 protein, rather than reduced intrinsic GAP activity.
AKT phosphorylation is known to be down-regulated in Tsc1/Tsc2 null cells due to a feedback circuit that leads to reduced expression and function of IRS-1 and PDGFR (14,15), as well as the effects of the TSC1/TSC2 complex on the activity of mTORC2, the AKT S473 kinase (16,27). To explore the basis of the down-regulation of AKT activation in the Tsc2del3/del3 MEFs, we examined levels of expression of IRS-1, PDGFRβ and PTEN, all potential mediators of AKT activation (Fig. 5C). None was significantly different in the Tsc2del3/del3 MEFs in comparison with controls.
To examine this feedback regulatory circuit downstream of TSC1/TSC2 in vivo, we also examined embryo lysates directly, derived from Tsc2del3/+ intercrosses collected at E13.5 (Fig. 5D). Reduction but not absence of Tsc2 expression was seen in the Tsc2del3/del3 embryos, as were higher levels of pS6K(T389). However, pS6 levels at both the S235/236 and S240/244 sites showed a minimal increase in comparison with controls (Fig. 5D and F). These findings are similar to those seen in the serum-stimulated MEF cultures (Fig. 5A, right), and likely reflect the active growth stimulation occurring during embryonic development, with high mTORC1 activation and thus minimal effect of loss of TSC1/TSC2 on this pathway.
Both pAKT(T308) and pAKT(S473) levels were markedly reduced in the embryo lysates, as were levels of pGSK3β(S21/9), a downstream substrate and effector of AKT. We did not see evidence of widespread apoptosis in the Tsc2del3/del3 embryos by either immunohistochemistry (Supplementary Material, Fig. S1) or immunoblotting (data not shown), suggesting that this reduction was not due to incipient embryonic death. These observations indicate that down-regulation of AKT activity is highly consistent in Tsc2del3/del3 embryos, and suggest the possibility that this effect, in addition to mTORC1 activation, contributes to fetal demise.
TSC is a distinctive familial hamartoma syndrome in which pathologically benign tumors cause significant morbidity and mortality through involvement of brain, kidneys and lungs (1). It is well known to be highly variable in severity, even among family members with the identical mutation, though in its classic form non-penetrance is not seen. In recent years, three missense mutations in TSC2 have been identified in patients from families that have a milder form of TSC disease, with some individuals bearing the TSC2 missense mutation that do not meet established TSC diagnostic criteria (9–11). The TSC2-R905Q protein produced by one of these mutations interacts normally with TSC1, and has partial function as a GAP for rheb in HEK293 assays (9). However, the activity of TSC2-R905Q has not been assessed when it is expressed at normal levels, as this is technically difficult to perform. Nonetheless, these clinical and biochemical data suggest that these are hypomorphic alleles of TSC2.
Here we present a detailed characterization of a novel mutation in Tsc2 in which there is deletion of the third coding exon, leading to deletion of 37 amino acids (residues 76–112, which show 92% identity between the human and mouse sequence) near the N terminus of Tsc2. The Tsc2-del3 mutant protein is expressed at much lower levels than wild-type Tsc2, binds to Tsc1, and has activity similar to that of wild-type Tsc2 when expressed at equivalent levels in HEK293 cells. However, both Tsc2del3/del3 MEF cell lines and embryos show marked activation of mTORC1, similar to that of Tsc2 null MEFs. Remarkably, the suppression of activation of AKT, as assessed by phosphorylation at the S473 site, seen in both the cultured Tsc2del3/del3 cells and in vivo in embryos is as marked as the activation of mTORC1 (Fig. 5A and D), highlighting the important role of TSC1/TSC2 in the regulation of AKT. Previous studies have identified the important effects of activation of mTORC1 in leading to reduced expression and function of the IRS and PDGFR proteins, through both phosphorylation and inactivation (IRS only) and transcriptional mechanisms (both IRS and PDGFR) (14,15,28,29). In the Tsc2del3/del3 cell lines and embryo lysates, expression of IRS and PDGFR does not appear to be altered (Fig. 5C and D). Recent observations (16) have indicated the importance of TSC1/TSC2 as a cofactor for the kinase activity of mTORC2, the S473-AKT kinase (27,30). Our data suggest that this is the predominant mechanism of suppression of AKT activation in the Tsc2del3/del3 cells and embryos. Thus, these observations highlight the consistency of downregulation of AKT activity in cells with mutational inactivation of TSC1/TSC2, even in the presence of limited residual functional TSC1/TSC2.
Pathological studies demonstrated that, similar to previous observations in Tsc1 null and Tsc2 null embryos (17,20), there is abnormal development of the liver and its hematopoietic islands in the Tsc2del3/del3 embryos. However, in addition, there were consistent and dramatic alterations seen in the vascular system in association with hemorrhage in multiple sites. These effects may be secondary to deficient hematopoiesis. However, they also raise the possibility that TSC1/TSC2 has a particular role in vascular development, a possibility that matches the frequency of vascular and smooth muscle cell lesions seen in both Tsc mouse models and TSC patients. This effect may also be due to either mTORC1 activation or AKT down-regulation, as AKT is known to play an important role in vascular system development (31).
It is interesting that the Tsc2del3/+ mice develop renal tumors at a considerably lower rate than Tsc2+/− mice. This occurs in the setting of slightly reduced mTORC1 activation in Tsc2del3/del3 cells, in comparison with Tsc2 null cells, and a similar slight reduction in the rate of suppression of AKT activity in Tsc2del3/del3 cells, in comparison with Tsc2 null cells. Previous considerations in both mouse models and patients have suggested that AKT suppression may be an important mechanism that restrains the growth and/or conversion to malignancy of TSC tumors (12–15). The current observations on this hypomorphic allele suggest that it is mTORC1 activation that is the primary determinant of growth by renal tumors in these mice, rather than growth being restrained by AKT suppression.
The in vivo hypomorphic function of the del3 allele, in contrast to the null alleles of Tsc2 previously generated, confirms that hypomorphic alleles at the TSC1 and TSC2 genes exist, and have phenotypes that are intermediate in nature. Another hypomorphic allele of Tsc2 in the mouse has also been described, made in the course of construction of a conditional allele of Tsc2, in which a neo cassette was inserted into the first intron of Tsc2 (22). In limited studies, Tsc2neo/neo embryos were reported to survive out to E17 in some cases (22). Only renal cysts, and no renal tumors were seen in Tsc2+/neo mice at up to 20 months of age. Surprisingly, immunoblotting showed near absence of Tsc2 protein in a Tsc2neo/neo embryo (22).
Although the majority of TSC1 and TSC2 mutations identified in TSC patients are completely inactivating, 10–20% may be hypomorphic alleles, including missense mutations, inframe indels and splice site mutations, all of which may have residual partial function. This is one important mechanism by which variability in the TSC patient phenotype may occur. Clearly there are many others, including stochastic variation in the frequency of second hit events during development. One important consequence of this observation, however, is the possibility that low dose rapamycin therapy which would attenuate though not eliminate mTORC1 activity, might mimic the effects of a hypomorphic allele of TSC2, and lead to reduced TSC hamartoma progression. Full dose rapamycin was recently reported to have benefit for the both the renal and lung lesions of TSC, but with moderate toxicity such that 20% of patients discontinued the treatment (32). Thus, low dose rapamycin administered over the long term (years, decades) might be similar to a hypomorphic allele of TSC2 in which growth rate is extremely slow, with minimal toxicity.
A modified targeted allele of Tsc2 in the mouse was generated using methods described previously (20). Briefly, genomic fragments from the Tsc2 allele (17) were assembled into a targeting construct, as shown in Fig. 1A. The construct was electroporated into J1 ES cells, which were then selected using G418. Ninety-six colonies were minimally expanded, DNA prepared and digested for analysis by Southern blotting to identify ES clones that had undergone homologous recombination. One such ES clone was subjected to transient cre expression to derive subclones in which the neo cassette had been eliminated by recombination, also identified by Southern blotting. One such ES subclone was analyzed by cytogenetics to confirm retention of diploidy, and injected into blastocysts, followed by transfer to pseudo-pregnant female mice. Multiple chimeric offspring were obtained, and were bred with wild-type C57BL/6 to found a conditional allele colony. Mice bearing the conditional allele were subsequently bred to a nestin-cre allele (33) in which recombination occurs in the germline, to derive mice with the deleted allele.
All procedures were carried out in accordance with the Guide for the Humane Use and Care of Laboratory Animals, and the study was approved by the Animal Care and Use Committee of Children’s Hospital, Boston. Individual mice were euthanized when weight loss of 20%, greatly reduced movement or other signs of morbidity were seen.
For studies of embryos, timed matings were performed using the vaginal plug to assess fertilization.
DNA samples were prepared from mice by standard methods. Genotyping was performed by PCR under standard conditions. Primers used for genotyping the conditional del3 allele were F3112 GATGAAGGCAGACCGTCAC and R3263 GAGGCTCAGGAGCTACAGTGA, producing a wild-type product of size 152 bp, and conditional c-del3 product of size 202 bp. Primers used for genotyping the del3 allele were F2300 TGATGAGGCAGTTTTTACCATTT, F3112 and R3263, producing a wild-type product of size 152 bp and del3 product of size 254 bp.
RT–PCR was performed to assess the relative level of Tsc2 mRNA. RNA was extracted from MEF lines using RNAeasy (Qiagen), and was reverse transcribed using Superscript III RT (Invitrogen, Carlsbad, CA, USA) and random hexamers. Quantitative PCR analysis was performed with primers for 18S rRNA and murine Tsc2, using SYBR green on an ABI7500 (Applied Biosystems).
Standard histology sections of mouse kidneys were prepared by rapid removal and 10% formalin fixation at 4°C with cutting into five 1–2 mm sections. Following paraffin embedding, 10 µm sections were cut and stained by H&E. Both gross and microscopic kidney pathology was read by a blinded observer (KP) and scored according to a modification of a formula used previously (34). The kidney tumor score for kidney cystadenomas was determined as a summed score for all lesions in a kidney, scoring each individual tumor grossly as follows: 1 for tumors < 1 mm; 2 for 1–1.5 mm; 5 for 1.5–2 mm; 10 for > 2 mm. Microscopic kidney tumor scores were determined similarly, except that the score for each lesion was multiplied by two if the tumor had a papillary component, and by four if it was a solid adenoma. The percent cellularity of each cystadenoma was determined as 0% for a cyst, 100% for a solid adenoma and 10–90% for cysts with papillary projections according to the degree of filling of each lesion. Comparison between sets of mice for tumor measurements was made using the non-parametric Mann–Whitney test.
Embryos were prepared for standard histology using Bouins fixative, paraffin embedding, cutting of serial 10 µm sections and H&E staining.
Antibodies used were: TSC1, p-p70S6K(T389), pAKT(Thr308), pGSK3α/β(S21/9), S6, pS6(S235/236), pS6(S240/244) from Cell Signaling Technology, Bedford, MA, USA; Tsc2 (C20), AKT1 (C20), Erk2 (K23), endomucin (V.7C7) from Santa Cruz Biotechnology, Santa Cruz, CA, USA; PTEN, PDGFRβ, IRS-1, Nestin, NeuN, MBP, GAPDH from Millipore, Billerica, MA, USA; GSK-3β from BD Biosciences, Franklin Lakes, NJ, USA; cyclin D1 from Thermo Scientific, Fremont, CA, USA; pAKT(S473) clone 14-5 from DAKO S/A, Denmark; Ki-67 from Vector Laboratories, Burlingame, CA, USA; Flag, HA from Sigma-Aldrich, St Louis, MO, USA.
For immunohistochemistry of both embryos and kidneys, tissues were rapidly removed, fixed overnight at 4°C in 4% paraformaldehyde, and then placed in 30% sucrose at 4°C for 1–2 days. Ten micrometer sections were cut on a cryostat (Leica, Germany), placed on slides and stained by the immunoperoxidase technique using DAB, with counterstaining with hematoxylin. For immunofluorescence microscopy, similar procedures were used, but using secondary antibodies labeled with Alexa Fluor 568 or Oregon green 488, and viewed on a Nikon (Tokyo, Japan) TE2000-E inverted microscope (35).
Apoptosis was assessed using the ApopTag Peroxidase In Situ-Apoptosis Detection Kit (Millipore, Temecula, CA, USA), using methyl green (Vector Laboratories) as counterstain.
MEF cultures were established from E10.5 to 13.5 embryos as previously described (15). These were maintained in DMEM with 10% fetal calf serum (FCS) in a humidified chamber with 5% CO2. Cells were passaged following a 3T3 protocol, as done previously. Immortalization was determined by recurrence of active cell growth that persisted beyond P30, after a period of senescence. For analysis of signaling pathway activation, cell lysates were prepared from cells that had been either starved of serum for 24–48 h or similarly starved and then treated with 10% FCS for 30 min, prior to preparation of lysates.
Cell lysates for immunoprecipitation were prepared in a lysis buffer including 1% NP-40, 1 mm PMSF, 1 mm Na3VO4, 1 mm NaF, 10 mm β-glycerophosphate and Protease Inhibitor cocktail (Sigma P8340). After centrifugation at 14 000g for 5 min at 4°C, extracts were precleared with Protein A/G PLUS-Agarose beads (Santa Cruz, sc-2003). Protein concentrations were normalized to 500–600 µg per sample, and incubated with 1 µg antibody at 4°C overnight, then 20 µl Protein A/G PLUS-Agarose beads were added and mixed for 1–2 h, beads were washed 5–10 times with extraction buffer, then treated with Laemmlli buffer, and eluates loaded on polyacrylamide gels.
Human embryonic kidney HEK293T cells were cultured in DMEM with 10% FCS. Transient transfections were performed using Lipofectamine2000 (Invitrogen). Full-length murine Tsc2 and Tsc2-del3 cDNAs were cloned into pRK7. HEK293T transfection mixes contained 7 µg of total DNA per well (of a six-well plate): 1 µg of HA-S6K1 in pRK7, 2 µg of Flag-TSC1 in pRK7 and 4 µg of some combination of murine Tsc2, Tsc2-del3 or empty pRK7.
Immunoblotting was performed as described previously (35). Each immunoblot lane shown is from the same gel/blot/film exposure. Equal protein loading for each well on each gel was determined by Bradford or DC Protein assays (Bio-Rad, Hercules, CA, USA), and confirmed by Coomassie blue staining of a parallel gel not used for immunoblotting. Cycloheximide was used as a protein synthesis inhibitor at 1 mm. Densitometry was performed using an Epson 1680 scanner.
National Institutes of Health (2R37NS031535 to D.J.K.).
We thank Rod Bronson for assistance with interpretation of mouse pathology; Lynsey Meikle for assistance with immunoblotting; Brendan Manning for the gift of Flag-TSC1 and HA-S6K1 expression constructs.
Conflict of Interest statement. None declared.