Hutchinson-Gilford progeria syndrome (HGPS) is a rare human genetic disorder characterized by striking progeroid features. Clinical findings in the skin include scleroderma, alopecia and loss of subcutaneous fat. HGPS is usually caused by a dominant-negative mutation in LMNA, a gene that encodes two major proteins of the inner nuclear lamina: lamin A and lamin C. We have generated tetracycline-inducible transgenic lines that carry a minigene of human LMNA under the control of a tet-operon. Two mouse lines were created: one carrying the wild-type sequence of LMNA and the other carrying the most common HGPS mutation. Targeted expression of the HGPS mutation in keratin-5-expressing tissues led to abnormalities in the skin and teeth, including fibrosis, loss of hypodermal adipocytes, structural defects in the hair follicles and sebaceous glands, and abnormal incisors. The severity of the defects was related to the level of expression of the transgene in different mouse lines. These transgenic mice appear to be good models for studies of the molecular mechanisms of skin abnormalities in HGPS and other related disorders.
Progeroid disorders can provide valuable insights into the genetic mechanisms underlying aging (Bohr, 2002; Martin and Oshima, 2000). Hutchinson-Gilford progeria syndrome (HGPS, progeria) is a rare genetic disorder affecting children with features suggestive of premature or accelerated aging. HGPS children appear normal at birth, but begin to display features of the disease within their first year(s) of life. Affected children experience delayed growth, are short in stature, have universal alopecia and suffer from restricted joint mobility and osteoporosis. Other key abnormalities include prominent scalp veins, delayed eruption of teeth, impaired sexual maturation, and a thin and high-pitched voice (DeBusk, 1972). Most children die in their early teens from heart attack or stroke, attributable to rapidly progressive atherosclerosis (see the Progeria Research Foundation's medical and research database at www.progeriaresearch.org).
The skin of children with progeria is described as having an aged look. Consistent progressive clinical findings include atrophic epidermis, dermal fibrosis (scleroderma-like with thickening, hyalinization and disorganization of collagen bundles), thin or absent hypodermis and a complete loss of skin appendages or decrease in number of hair follicles and sebaceous glands (Ackerman and Gilbert-Barness, 2002; DeBusk, 1972; Erdem et al., 1994; Gillar et al., 1991; Hutchison et al., 2001; Jansen and Romiti, 2000; Rodriguez et al., 1999; Sevenants et al., 2005; Stables and Morley, 1994; Plasilova et al., 2004) (the GENEReviews' database at www.geneclinics.org). Hypoplastic eccrine glands and spotty skin pigmentation have also been reported (DeBusk, 1972; Jansen and Romiti, 2000).
The inheritance pattern of HGPS is autosomal dominant. It is reproductive lethal. At least 90% of the cases are due to a de novo mutation in a single nucleotide in exon 11 of the LMNA gene, 1824C>T (G608G). The mutation partially activates a cryptic splice site and produces a mRNA that codes for lamin A protein with an internal deletion of 50 amino acids (De Sandre-Giovannoli et al., 2003; Eriksson et al., 2003). The LMNA gene encodes lamin A, lamin C, lamin AΔ10 and lamin C2 (Burke and Stewart, 2002; Fisher et al., 1986). Lamin A and lamin C are major proteins of the inner nuclear lamina, located beneath the nuclear envelope. The lamina is believed to give the nucleus its shape and strength and to play significant roles in DNA replication (Burke and Stewart, 2002; Moir and Spann, 2001). The inner nuclear lamina is also essential in defining higher order structure by providing anchoring sites for chromatin domains and various proteins at the nuclear periphery (Burke and Stewart, 2002). Lamin A is synthesized as a precursor protein, prelamin A, which is rapidly subjected to posttranslational processing to produce mature lamin A. The processing is initiated by farnesylation of the C-terminal CAAX motif (CSIM), removal of the last three amino acids (–SIM) and methyl esterification of the cysteine. The final step to yield mature lamin A, cleavage by the metalloproteinase ZMPSTE24, discards the 15 C-terminal residues of the prelamin A molecule previously inserted in the inner nuclear lamina (Bergo et al., 2002; Hutchison et al., 2001; Sasseville and Raymond, 1995; Stuurman et al., 1998). In progeria, the exon 11 deletion removes 50 amino acids, including the recognition site for ZMPSTE24, and therefore progeria children accumulate a indigestible farnesylated prelamin A molecule (referred to as progerin or LAdel50). Without the ability to be released from its lipid tether, progerin apparently interacts aberrantly within the nuclear lamina, interfering with its structure, intranuclear architecture and macro-molecular interactions, and collectively producing major impact on nuclear functions. The presence of progerin has been shown to lead to lobulated nuclei, thickening of the lamina, loss of peripheral heterochromatin, clustering of nuclear pores and derangement of normal mitosis (Cao et al., 2007; Dechat et al., 2007; Eriksson et al., 2003; Goldman et al., 2004).
There are at least nine different autosomal recessive and dominant genetic diseases linked to mutations in the LMNA gene, collectively called laminopathies (Capell and Collins, 2006; Worman and Courvalin, 2004). Several mouse models, particular useful for studies on laminopathies, have been published (Bergo et al., 2002; Mounkes et al., 2003; Pendas et al., 2002; Sullivan et al., 1999; Varga et al., 2006; Yang et al., 2006). Mice homozygous for the Emery-Dreifuss muscular dystrophy point mutation, L530P, and defective splicing of lamin A and lamin C transcripts, develop severe growth retardation and die within 4 to 5 weeks. These mice have a slight waddling gait, small jaws, abnormal dentition, thickened epidermal layer with regions of hyperkeratosis, thinning of dermis, absence of subcutaneous fat and osteoporosis (Mounkes et al., 2003). A BAC transgenic mouse that carries the G608G mutated human LMNA shows no external phenotype, but demonstrates progressive abnormalities of large arteries that closely resemble the most lethal aspect of the human phenotype (Varga et al., 2006). A recent report from Yang and co-workers presents evidence of severe growth retardation, bone disease and a reduction of subcutaneous fat in a knock-in mouse model of progeria (Yang et al., 2005; Yang et al., 2006).
In the present study we describe the development of inducible LMNA transgenic mouse lines based on the tet-on/off system (Gossen and Bujard, 1992; Zhu et al., 2002). This tissue-specific expression system is particularly useful for the study of gene products that might be toxic when expressed early during development, or products that could have a negative effect on reproduction. We constructed two tet-operon driven transgenic lines, one containing a minigene of wild-type human LMNA, and the other carrying the most common HGPS mutation (1824C>T). By breeding with K5tTA mice (Diamond et al., 2000), we obtain an inducible system that expresses human lamin A (LA) and progerin (LAdel50) in epidermal keratinocytes. In this study, we show that expression of the mutant LMNA allele in postnatal epidermis replicates several features of the HGPS skin phenotype.
Generation of tetop-LAwt and tetop-LAG608G transgenic mice
Minigenes of human Lamin A (LAwt and LAG608G) were constructed by PCR amplification of human genomic DNA and cDNA, subsequently digested with a unique restriction site in exon 11 (DrdI), purified and ligated to a tetop vector (gift of P. Scacheri). The human LA minigenes contained the complete coding region of lamin A, including exon 1-11, intron 11 and exon 12 (downstream of the stop codon) (Fig. 1). The tetop-LAwt and tetop-LAG608G transgenic lines were created by injecting a fragment of 4165 bp containing the lamin A minigene, and the upstream tet-operon (tetop), downstream internal ribosomal entry site (IRES), the coding region for eGFP and a SV40 poly A site. Twenty founders of each minigene, as determined by PCR genotyping, were born and bred to obtain F1 lines.
Screening for minigene expression and expression regulated by the K5tTA
F1 animals were intercrossed to K5tTA transgenic mice (Fig. 1C), and minigene expression from different F1 lines was evaluated by western blot and immunofluorescence (Figs 2, 3). Skin biopsies from single transgenic tetop-LAwt and tetop-LAG608G littermates were also analyzed by western blot to confirm protein expression from the minigene only in the presence of the K5 transactivator. Bi-transgenic offspring were obtained from 11 F1 lines (from seven founders) of tetop-LAwt and eight F1 lines (from six founders) of tetop-LAG608G. Expression of GFP was detected in bi-transgenic offspring from eight of the 11 F1 lines of tetop-LAwt, and expression in five of the GFP positive F1 lines demonstrated tight regulation by the tet-operon. The remaining three F1 lines showed expression in non-K5tTA containing tissues and were discarded from the study (data not shown). Expression of GFP was detected in bi-transgenic offspring from seven out of the eight F1 lines of tetop-LAG608G. One tetop-LAG608G+; K5tTA– F1 line was excluded owing to GFP expression in protein extracts in the absence of the transactivator (data not shown). Minigene expression [human LA and LAdel50 (progerin)], were also screened by western blot. LA was present in four bi-transgenic F1 lines of tetop-LAwt (SF1-04, SF1-02, EF1-02 and EF1-03) (Fig. 2B, data not shown). LA and LAdel50 were present in three bi-transgenic F1 lines of tetop-LAG608G (VF1-07, CF1-05 and DF1-03) (Fig. 2B, data not shown). To evaluate the transgenic expression pattern in the skin, we performed immunofluorescence with an antibody specific for human lamin A/C on skin sections from all F1 lines that were positive for GFP on western blot, and that did not show expression in single transgenic animals (Fig. 3). Minigene expression was detected in the hair follicle and the interfollicular epidermis in one F1 line (VF1-07) of tetop-LAG608G (Fig. 3A-I). Bi-transgenic animals of the other F1 lines show expression in only a few cells of the hair follicle, although the same F1 lines were previously positive by western blot with human lamin A/C antibody (CF1-05 and DF1-03) (data not shown). Lamin A minigene expression was detected in the hair follicle and the interfollicular epidermis in two F1 lines [SF1-04 and SF1-02 (from the same founder, S)] of tetop-LAwt (Fig. 3M-O). The other three F1 lines showed expression in only a few cells mainly of the hair follicle (EF1-02 and EF1-03 from founder E) or no expression (data not shown), confirming the previous results from the western blots.
Other tissues known to contain keratin 5-expressing cells were also investigated for minigene expression. Immunofluorescence with human lamin A/C antibody was performed on sections from esophagus, salivary glands, stomach and tongue of 9-week-old bi-transgenic animals of F1 line VF1-07, tetop-LAG608G. Expression was detected in myoepithelial cells of the salivary gland, basal and suprabasal cells of esophagus, stomach and tongue, indicating correct targeting of the transgene to keratin 5-expressing tissue (supplementary material Fig. S1, data not shown). No minigene expression was detected in tissues from tetop-LAG608G+; K5tTA– transgenic mice, indicating tight regulation from the tet-operon (data not shown). In mice receiving doxycycline in their drinking water, no minigene expression was seen in dorsal skin sections. Minigene expression was first noted in dorsal skin sections 7 days post-doxycycline removal.
Transgenic expression, levels and prelamin A accumulation
PCR with primers specific for human LA and lamin Adel150 (LAdel150) was performed on cDNA from skin of bi-transgenic and control mice from F1 lines (Fig. 2A, data not shown). Amplification of human LA and progerin (LAdel150) gives two fragments of 276 bp and 123 bp, respectively (Fig. 2A, lanes 1, 2). All bi-transgenic animals from the selected F1 lines have a fragment corresponding in size to human LA (Fig. 2A, lanes 4-7, 9-11), indicating correct splicing of the minigene. In bi-transgenic animals from F1 lines of tetop-LAG608G, there is also an additional fragment corresponding in size to progerin (Fig. 2A, lanes 4-7), indicating that the 1824C>T mutation is recognized by the mouse splicing apparatus. Protein extracts from skin of bi-transgenic and control animals of different F1 lines were analyzed by western blot. Approximately equal levels of LA transgenic expression were identified in bi-transgenic animals of F1 lines VF1-07 and SF1-04 when compared with β-actin (Fig. 2B). Lower levels of minigene expression were seen in bi-transgenic animals from F1 lines EF1-03, DF1-03 and CF1-05 (Fig. 2B, data not shown), which correlated with the expression seen on immunofluorescence.
Relative levels of human LA to mouse lamin A were quantified using densitometry on western filters hybridized with an antibody that recognize both human and mouse lamin A/C (Fig. 2C, data not shown). The average overexpression of lamin A in dorsal skin samples from bi-transgenic animals of VF1-07 was 0.88 and for SF1-04 it was 0.73.
A protein corresponding in size to prelamin A was identified in lanes with protein extracts from bi-transgenic F1 lines, VF1-07 and SF1-04 using the human lamin A/C and the prelamin A antibodies [Fig. 2C (right filter), data not shown]. Additional western experiments with increased separation between lamin A and prelamin A confirmed the presence of prelamin A accumulation in both lines [Fig. 2C (left filter)].
Skin histopathology of F1 line VF1-07, tetop-LAG608G+; K5tTA+ transgenic mice
Tetop-LAG608G, F1 line VF1-07, transgenic mice were intercrossed to K5tTA transgenic mice. Doxycycline was removed on postnatal day 21 (dox day 21). Six weeks after doxycycline removal, skin abnormalities were apparent (Fig. 4B-G). At this stage, lesions were in patches of varying severity (Fig. 4B). The less affected regions were characterized by a slight to moderate hyperplasia of the interfollicular epidermis associated with hypergranulosis and hyperkeratosis. There were dystrophic changes in hair follicles while the associated sebaceous glands were beginning to show hyperplasia and irregular maturation of the sebocytes [Fig. 4B (right), D,G]. At this stage, there were increased numbers of inflammatory cells, including polymorphonuclear granulocytes in the dermis. In the more severely affected regions, there were papillary changes with severe epidermal hyperplasia, hyperparakeratosis and enlargement and displacement of sebaceous glands [Fig. 4B (left part), C-F]. The structure of the sebaceous glands was immature and did not appear to be fully differentiated (Fig. 4F,I). Dermal changes included fibrosis and moderate to severe inflammatory cell invasion. The end-stage, within 17 weeks of turn-on of expression of the transgene, was characterized by loss of hypodermis, fibrosis of the dermis, hypoplastic sebaceous glands and small hyperchromatic nuclei oriented in parallel with the basement membrane; the epithelium still shows stratification with maturation of the cornified layer (Fig. 4J). Dorsal skin of F1 line of VF1-07, tetop-LAG608G+; K5tTA–, showed a normal structure of the skin up to 1 year and 9 months (Fig. 4A,K, data not shown).
Skin phenotype in the tetop-LAG608G mouse lines expressing low levels of the transgenes.
Dorsal skin of bi-transgenic animals for F1 line CF1-05 of tetop-LAG608G is normal and, even at 1 year and 9 months of age, is indistinguishable from wild-type controls (data not shown). The dorsal skin of bi-transgenic animals for tetop-LAG608G F1 line DF1-03, show a possible slight hypoplasia of the epidermis, but otherwise normal structure of the other layers of the skin at 2 years of age (data not shown). The phenotypes of these mouse lines suggest that sustained lower expression levels of lamin A and progerin (within only a few cells of the hair follicle), does not have any significant effect on the structure of the epidermis.
Skin histopathology of tetop-LAwt+; K5tTA+ transgenic mice
The mouse lines overexpressing wild-type human LA were created as a control for our experiments. Histopathology of the skin from bi-transgenic tetop-LAwt F1 lines SF1-04, SF1-02, EF1-03 and EF1-02 were analyzed for histopathology. No significant skin lesions were noted in bi-transgenic animals of F1 lines SF1-04, EF1-02 or EF1-03 older than 1 year of age (Fig. 4L, data not shown). Bi-transgenic animals of F1 line SF1-02 (dox day 21) show a slight to moderate epidermal hyperplasia and hyperparakeratosis, with hyperplasia of the sebaceous glands, and fibrotic dermis, already present at 6 weeks (data not shown). However, higher levels of LA (average human LA relative to mouse lamin A were 1.5) and significant amounts of progerin were detected in this line of mice (data not shown), and makes it hard to interpret. In the single transgenic animals for tetop-LAwt, of the same F1 lines skin structure is normal. Taken together, these data suggests that overexpression of tetop-LAwt in the skin is not resulting in any pathological changes.
Hyperproliferation in dorsal skin of tetop-LAG608G+; K5tTA+ transgenic mice
Immunohistochemistry with cytokeratin 5 and cytokeratin 6 antibodies on sections from dorsal skin tetop-LAG608G+; K5tTA+ F1 line VF1-07 with severe epidermal hyperplasia show mislocalized keratin 5 and keratin 6 expression (Fig. 5C-E). Anti-phospho-histone H3 (a marker for hyperproliferation) showed increased proliferation especially in regions of severe epidermal hyperplasia (Fig. 5G,H). Increased apoptosis, measured with cleaved caspase 3 antibodies, was not seen in these areas or in sections from end-stage disease of the same transgenic line (data not shown).
Epidermal differentiation in dorsal skin of tetop-LAG608G+; K5tTA+ transgenic mice
We used immunofluorescence with antibodies against keratin 1, keratin 10, filaggrin and loricrin to examine whether epidermal differentiation was altered in regions of epidermal hyperplasia in bi-transgenic animals of the F1 line VF1-07. Similar to wild-type skin, keratin 1- and keratin 10-positive keratinocytes resided in the suprabasal layers of bi-transgenic epidermis (Fig. 6A,B,D,E). Although the total expression of keratin 1 and keratin 10 was increased in regions of epidermal hyperplasia, the majority of the expression was seen in the spinous layers and only a few cells of the basal layer were seen to express keratin 1 and keratin 10 (Fig. 6C,F). Immunostaining using antibodies to filaggrin and loricrin showed expression in the granular layer of wild-type and bi-transgenic epidermis (Fig. 6G-L). Similar to the spinous markers, filaggrin expression was seen in multiple layers of the thickened epidermis (Fig. 6I). Filaggrin-expressing cells were also seen in the dermis of bi-transgenic skin (Fig. 6H,I).
Premature death and external phenotype
Hair-thinning, growth retardation and premature death were noted in bi-transgenic animals from F1 line VF1-07 (tetop-LAG608G+; K5tTA+) (Fig. 7). Intercross breeding pairs were supplied with doxycycline, which was removed at weaning, postnatal day 21 (dox day 21) or postnatal day 0 (dox day 0). Hair follicle density was measured by counting hair follicles in sections from dorsal skin from bi-transgenic and wild-type mice. Although bi-transgenic mice experienced hair thinning on gross examination, there was no reduction in the number of hair follicles compared with wild-type mice. Weights were recorded weekly starting at postnatal week 1. Growth retardation and early death were noted in bi-transgenic animals with a median survival of 14 weeks for dox day 21 (Fig. 7A,D) and median survival of 7 weeks for dox day 0 (Fig. 7B,E). In an effort to identify causes of premature death in these animals, we identified dental problems (described below), which we suspected could have significant effect on their lack of growth. To test whether this was the case, we fed the mice a softer diet of dissolved pellets on the cage floor. With this special feeding the median survival increased from 7 to 29 weeks (compare Fig. 7B,E with 7C,F). Weights were also recorded for bi-transgenic mice of F1 line DF1-03, of tetop-LAG608G, but were indistinguishable from control mice (data not shown). These mice did not show any external phenotype, except for a slight hair thinning at 2 years of age (data not shown). Weights were also monitored on bi-transgenic animals from F1 line SF1-04, tetop-LAwt (dox day 21), but no premature death or abnormal weights were seen for this line (+/+, n=9) (data not shown). However, all bi-transgenic animals of the SF1-04 line showed partial hair loss, and regions with crusting of the skin first noted at 7 weeks post dox removal (+/+, n=10) (data not shown). No premature death or significant changes in weight were noted in bi-transgenic animals of F1 line EF1-02 of tetop-LAwt (data not shown). Premature death with a median survival of 8.7 weeks, and partial hair loss with regions of skin crusting were also noted in bi-transgenic animals of F1 line SF1-02 (tetop-LAwt+; K5tTA+) (data not shown). Higher levels of LA (average relative to mouse lamin A were 1.5) and significant amounts of progerin were detected in this line of mice, which makes it hard to interpretate the phenotype (data not shown).
Histopathology of organs other than skin
We examined the gastrointestinal system, liver, pancreas, spleen, thymus, mammary glands, salivary glands, tear glands, brown fat, kidneys, adrenal glands, reproductive organs, skull, tongue, brain, the respiratory mucosa of the nasal cavity, trachea and lungs on bi-transgenic animals of F1 line VF1-07 (tetop-LAG608G) and F1 line SF1-04 (tetop-LAwt) of different ages. Bi-transgenic animals (with dox stopped at day 21) of F1 line VF1-07 were sacrificed for histopathology at 9 weeks (n=3) and at 17-20 weeks (n=4), and bi-transgenic animals of F1 line SF1-04 were sacrificed at 22-24 weeks (n=3). In bi-transgenic mice from F1 line VF1-07, the only significant recurrent changes included a possibly mild hyperplasia in the stomach and more extensive changes involving both the lower incisors and surrounding tissues. The latter changes ranged from increased impaction of food or bedding material in the gingival sulcus with associated acute inflammatory reaction to overt necrosis of the pulp. In the latter cases, foreign material with the appearance of cellulose was found in the pulp (supplementary material Fig. S2B). Inflammation spread into the periodontal ligament and surrounding bony structures but no changes were seen in the upper jaws. No significant recurrent changes were observed in bi-transgenic animals of tetop-LAwt. Single transgenic littermates for the respective minigene of different ages, were also analyzed (n=7), and no significant abnormalities were observed in these mice.
For diseases like HGPS that are both devastating and rare, animal models are essential to understand the molecular basis of the disease and investigate what kinds of interventions may benefit patients. Transgenic and knockout mice have been used to shed light on the genetics behind laminopathies and aging in general. In the transgenic mouse model described here, we decided to leave the mouse Lmna gene intact and to generate transgenic mice that express minigenes of wild-type and HGPS human LA under the control of an inducible promotor, providing the opportunity to limit expression to a specific tissue. Using the tissue specific tet-on/off system, we developed two lines carrying minigenes of human LA, differing only in nucleotide 1824, where one transgene carries the wild-type sequence of LMNA, and the other carries the most common HGPS mutation (G608G, 1824C>T). The clinical picture in HGPS involves multiple organs and tissues, and skin is one of the first organs to show typical signs of disease, including scleroderma, loss of subcutaneous fat and alopecia. Based on this, we decided to use the K5tTA (Diamond et al., 2000) as the transactivator, which is predicted to direct the expression of the minigenes to basal cells of the interfollicular epidermis and hair follicle. Screening multiple bi-transgenic offspring from multiple F1 lines from different founder animals identified lines that showed different levels and pattern of expression, but only bi-transgenic animals from the F1 line VF1-07 of the tetop-LAG608G, and SF1-04 of tetop-LAwt, showed expression in cells of the interfollicular epidermis and hair follicle. Accurate intron splicing and partial splice site activation of the 1824C>T mutation in the minigenes was demonstrated by RT-PCR, and confirmed by direct sequencing (data not shown). Protein expression of human LA and LAdel50 was detected with western blot and immunofluorescence in keratin 5-expressing tissues. The VF1-07 and SF1-04 lines show similarly high levels of human LA protein, whereas bi-transgenic animals of other F1 lines have lower expression levels. In bi-transgenic animals expressing progerin, there was a progressive phenotype beginning with regions of mild to severe epidermal hyperplasia, hyperparakeratosis and hyperplasia of the sebaceous glands, and progressing to an end stage characterized by loss of subcutaneous fat, fibrosis of dermis and hypoplastic sebaceous glands.
We are unsure what causes the observed variation in phenotype within the intermediate stage (with severe and mild regions even within the same section, Fig. 4B). However, immunofluorescence using the anti-human lamin A/C antibody on sections with regions of phenotypic variability showed staining of cells of the interfollicular epidermis and the hair follicle of both mildly and more severely affected regions, which excludes the possibility of mosaic expression of the transgene (data not shown). Our progerin-expressing mice also had an external phenotype with hair-thinning, low weight and premature death, with a median survival of 7 or 14 weeks, depending on when the transgene was turned on. Severe abnormalities include a well-developed dermal fibrosis, absence of detectable hypodermal adipocytes and premature death, all clinical features that are found in children with HGPS. Immunohistochemistry with antibodies to keratin 5, keratin 6 and phospho-histone H3 show mislocalized expression of keratin 5 and 6, with increased proliferation in regions of severe epidermal hyperplasia. This is in agreement with prior studies on mouse skin were suprabasal expression of keratin 5 and expression of keratin 6 in the interfollicular epidermis associates with hyperproliferative disease (Fuchs, 1995; HogenEsch et al., 1999). Analysis of the expression pattern of additional markers for epidermal differentiation (keratin 1, keratin 10, loricrin and filaggrin) indicated that regions with epidermal hyperplasia and increased proliferation contained layers of differentiated keratinocytes. Although there was an increase in the thickness of the spinous and granular layers compared with wild-type skin, the terminal differentiation of keratinocytes was essentially similar to what was seen in normal skin. Hyperproliferation and apoptosis has been reported in HGPS fibroblasts previously (Bridger and Kill, 2004). However, we did not detect any increased apoptosis in skin from these animals (data not shown).
Considering that nutritional problems might contribute to growth failures and early death, we looked into the possibility that these mice had problems with eating. The stomach of a few animals showed presence of mild hyperkeratosis (data not shown) but malnutrition was most probably due to dental abnormalities (inflammation and foreign material present within the pulp of the teeth) in the lower jaws of these mice. Feeding the animals a softer diet increased their median survival (from 7 to 29 weeks). Even though complete necropsies were performed and corresponding organs examined histopathologically, other potential contributing changes to the early mortality were not found, even in internal cells and organs known to express keratin 5.
Bi-transgenic mice of the SF1-04 line, with an average overexpression of human LA of 0.73, were used as a controls. Animals in the SF1-04 line did not show any weight loss, premature death or any histopathological abnormalities of the skin or other organs that were investigated, though all animals had an external phenotype of partial hair loss, with regions of skin crusting (not apparent on histopathology). It is interesting to note that a modest amount of prelamin A accumulates in this line, presumably owing to high levels of expression. But the mildness of the phenotype argues against major toxicity of this amount of prelamin A accumulation in the skin. This contrasts with the results reported in the Zmpste24-deficient mice (Bergo et al., 2002; Fong et al., 2004; Pendas et al., 2002), where virtually all of the mouse lamin A remains as prelamin A.
Surprisingly, bi-transgenic mice of an additional F1 line, SF1-02, of tetop-LAwt, with an higher overexpression of normal human LA (average relative to mouse lamin A of 1.5 compared with 0.73 for SF1-04), were also found to express a significant amount of progerin (data not shown). Prior studies have shown that progerin splicing occurs at a low level in human wild-type lamin A, with the G608G mutation greatly increasing the usage of this cryptic splice donor (Cao et al., 2007; Scaffidi and Misteli, 2006). For reasons that are not clear, this particular line used that splice donor quite efficiently, despite the wild-type sequence. These mice showed premature death, partial hair loss, skin crusting and some histopathological abnormalities of the dorsal skin similar to those present in bi-transgenic mice from the VF1-07 line (data not shown).
In summary, this study shows that expression of progerin in the skin of transgenic mice induces a progressive disease that begins with hyperproliferation and ends in a phenotype that resembles many of the clinical features of the skin reported in children with HGPS. This transgenic mouse model should therefore be useful in studying the pathogenesis and potential treatment of HGPS.
Materials and Methods
Generation of tetop-LAwt and tetop-LAG608G transgenic mice
The lamin A (LAwt and LAG608G) minigenes were constructed by PCR amplification of human genomic DNA (AG11498) (Eriksson et al., 2003) with primers 5′-GCTCTTCTGCCTCCAGTGTC-3′ and BamHI-GTCCCAGATTACATGATGCTG, and PCR amplification of human cDNA from an individual with HGPS (AG01972) (Eriksson et al., 2003) with primers EcoR1-ACTCCGAGCAGTCTCTGTCC and 5′-GGTCCCAGATTACATGATGCT-3′. Both PCR products were purified (QIAquick, Qiagen) and the genomic DNA PCR product were digested with DrdI and BamHI, and the cDNA PCR product were digested with DrdI and EcoR1. Following gel purification (Wizard SV Gel and PCR Clean-Up system, Promega), the LA minigene fragments were ligated to a vector that had been generated previously for another transgenic construct (tetop-Men1, a gift from P. Scacheri). The Men1 transgenic vector was constructed as follows: the complete coding sequence of the mouse Men1 gene was inserted into a tetracycline responsive cloning vector containing the mp-1 intron poly(A) sequence (gift of G. Fisher and H. Varmus, Sloan-Kettering). A 2.6 kb fragment containing the tet-operon and Men1 gene was PCR amplified from this construct and cloned into pIRES2-EGFP (Clontech). Prior to insertion of the LA minigene fragments, the Men1 gene was removed by an initial gel purification of a digest with BamHI, followed by a second gel purification of a digest with EcoR1. PCR screening of clones, with EcoR1-ACTCCGAGCAGTCTCTGTCC and BamHI-GTCCCAGATTACATGATGCTG, identified clones where the LA minigene had replaced the Men1 gene downstream of the tet-operon (tetop) and upstream of the eGFP. Sequencing of the clones with LMNA-specific primers identified minigenes of tetop-LAwt (wild-type sequence in codon 608) and tetop-LAG608G [G608G (1824C>T)]. Double CsCl banding (Lofstrand Laboratories, Gaithersburg, MD) was used to purify the vectors. The transgenic constructs (4165 bp) were released from its carrier by digests with AseI and NotI, gel purified and injected into the male pronucleus of recently fertilized Fvb/N embryos. The generation of tetop-LAG608G and tetop-LAwt transgenic mice was approved by the Animal care and Use Committee of the National Human Genome Research Institute and the National Institutes of Health, G-03-5. The animal studies were approved by the Stockholm South Ethical review board, Dnr. S148-03, S139-05, S111-05 and S141-06.
DNA was extracted from tail biopsies using standard phenol-chloroform protocol. Genotyping was performed with PCR for the LA minigenes (tetop-LAG608G and tetop-LAwt) (Eriksson et al., 2003) and K5tTA (Diamond et al., 2000). PCR with Myc primers was used as positive control for presence of DNA. Southern blotting was performed according to standard protocol using SacI. The probe was created by PCR amplification of human LMNA (exon 11, intron 11 and exon 12 to the stop codon) with primers 5′-ACCCCGCTGAGTACAACCT-3′ and ACATGATGCTGCAGTTCTGG-3′. The PCR product were TA cloned (TOPO TA-cloning kit, Invitrogen) and digested with EcoRI, to release a fragment of 609 bp that was gel purified and used as probe for the LA minigenes. For single transgenic integration, the probe hybridizes to a fragment larger than 3387 bp, depending on the next SacI site at the transgenic integration site. For multiple tandem integrations, the probe hybridizes to an additional fragment of 3690 bp. All filters contained a SacI digest of human genomic DNA and the probe hybridized to a fragment of 4449 bp. Estimation of transgenic copy number was made by comparison with SacI digests of non-transgenic genomic DNA spiked with different amounts of plasmid (1×, 5×, 10× and 20×) that contained the tetop-LAG608G transgenic construct, described above. The probe hybridizes to a fragment of 7034 bp in the spiked DNA. Calculations for estimation of transgenic copy number were in accordance with standard method (Gene Targeting and Transgenic Facility, University of Virginia Health System). F1 lines of tetop-LAwt: SF1-04, SF1-02 and EF1-03 each have an estimated copy number of 4. F1 lines of tetop-LAG608G: VF1-07, CF1-05 and DF1-03 have an estimated copy number of 1, 4 and 2, respectively (data not shown).
Animals were housed at the experimental unit animal housing facility at the Karolinska Hospital, Huddinge, Sweden. Animals were kept in open Makrolon 2 or 3 community cages, with a maximum number of five and 10 mice per cage, respectively. The housing conditions included a 12 hour light/dark cycle, a temperature of 20-22°C and 50-75% air humidity. Mice were supplied with R36 pellets (Lactamin, Sweden) and drinking water ad libitum.
Animal breeding and screening of transgenic lines
Colonies were maintained with mating to FVB/NCrl. During intercross breeding and up to postnatal day 0 or 21, the animals' drinking water was supplemented with 100 μg/ml doxycycline (Sigma) and 2.5% sucrose. The drinking bottles were covered in foil, and the water was changed every third day. Small skin biopsies were obtained from dorsal skin and/or tail skin of F1 and F2 that were positive for both the lamin A minigene and K5tTA (Diamond et al., 2000) (tetop-LAwt+; K5tTA+ and tetop-LAG608G+; K5tTA+). Skin biopsies were also obtained from tetop-LAwt+; K5tTA– and tetop-LAG608G+; K5tTA– transgenic mice, and served as negative controls for expression analysis. Weights were recorded weekly. In accordance with the Karolinska Institute guidelines for animal care, animals were sacrificed when a sudden weight drop was recorded (sudden weight loss of 25%) and/or if their general health status were compromised. Weight and survival curves, and statistical analysis were performed using the GraphPad Prism, version 4 software. Immunofluorescence with the human lamin A/C antibody showed that withdrawal of doxycycline from the drinking water resulted in transgenic expression first noted at day 7 post-doxycycline removal (n=7 skin samples).
Analysis of transgene expression (RNA and protein)
RNA was extracted from mouse tissues with Trizol or Micro-FastTrack 2.0 mRNA isolation kit (Invitrogen), using Lysing Matrix D and Fastprep 220A (Qbiogene). DNase treatment and cDNA synthesis were performed in accordance with previously published procedure (Eriksson et al., 2003). Expression of the LA minigenes were analyzed by PCR with 5′-AGTTCTGGGGGCTCTGGGT-3′, 5′-ACTGCAGCAGCTCGGGG-3′ and 5′-TCTGGGGGCTCTGGGC-3′. Expression of human LA gives a fragment of 276 bp, and expression of human LAdel150 gives a fragment of 123 bp, respectively. PCR with TBP-specific primers was performed on all samples as control (Eriksson et al., 2000).
Protein was extracted from mouse tissues in RIPA buffer or 8 M UREA, 5% RIPA (including a cocktail of proteinase inhibitors, Roche) and homogenized with Lysing Matrix D and Fastprep 220A (Qbiogene). Mini-gel western blots were performed in accordance with previously published procedures (Eriksson et al., 2003). A larger western blot system was used to enhance the separation of proteins (PROTEAN II xi Cell, BIORAD). Enhanced protein separation were accomplished using 1 mm thick 20 cm long 4%/7.5% discontinuous Laemmli slab gel. The gels were run for 12 hours at 25 mA using a cooling system at 2°C, according to the manufacturer's recommendation. Protein transfer was performed according to standard procedure for the Semidry Transfer Cell (BIORAD). Protein markers were BenchMark Pre-stained protein ladder (10748-101, Invitrogen) and PageRuler Prestained protein ladder (#SM0671, Fermentas). Primary antibodies used for western blot were: mouse monoclonal anti-human lamin A+C (mab3211, Chemicon), goat polyclonal anti-human prelamin A (sc-6214, Santa Cruz Biotechnology), goat polyclonal anti-lamin A/C (sc-6215, Santa Cruz Biotechnology), rabbit polyclonal anti-gfp (ab290, Abcam Ltd), mouse monoclonal anti-gfp (Sc-9996, Santa Cruz Biotechnology) and mouse monoclonal anti-β-actin (A5441, Sigma).
Protein quantification was performed on western filters, with enhanced protein separation, hybridized with an antibody to the N-terminal region of lamin A/C (sc-6215). The antibody recognized lamin A and C of both human and mouse origin. Relative band intensities of human LA to mouse lamin A in the same lane were quantified in protein extracts from dorsal skin samples from three bi-transgenic animals of F1 lines VF1-07, SF1-04 and SF1-02. The average relative levels were used to assess the degree of lamin A overexpression. Densitometry was performed using Versa Doc Imaging System (BIORAD) and analyzed using the Quantity One software.
Animals were sacrificed using an overdose of isoflurane and tissues were collected and fixed at +4°C overnight in 4% paraformaldehyde (pH 7.4). Following fixation, the tissues were dehydrated in ethanol and embedded in paraffin. All assays were performed on 4 μm sections. Sections were stained with Haematoxylin and Eosin according to standard procedures. Immunohistochemistry for cytokeratin 5, cytokeratin 6, phospho-histone H3 and cleaved caspase 3 were performed in accordance with previously published procedure (Svard et al., 2006). Immunohistochemistry for adipophilin (1:300; Progen), using a pressure cooker for antigen retrieval. Immunofluorescence experiments for human lamin A/C (1:30; mab3211, Chemicon), loricrin (1:500; PRB-145P, Covance), filaggrin (1:1000; PRB-417P, Covance), cytokeratin 1 (1:500; PRB-165P, Covance), cytokeratin 10 (1:500; PRB-159P, Covance) and cytokeratin 5 (1:1000; PRB-160P, Covance) were performed in accordance with previously described procedures (Varga et al., 2006).
Hair follicle density
Mice were intercrossed on doxycycline, which was removed at day of birth. From postnatal day 21 and onwards, the animals were supplied with a soft diet of dissolved pellets ad libitum. Animals were sacrificed at 7 weeks of age, and dorsal skin samples were collected and processed for Haematoxylin and Eosin staining. Number of follicles per mm section were counted in sections from two different anatomical dorsal skin regions (forehead and mid-dorsal region) in wild-type (n=3) and bi-transgenic animals of the VF1-07 line (n=3). Statistical analysis was performed using unpaired two-tailed Student's t-test (Prism, GraphPad).
We acknowledge the technical assistance of Jun Cheng, Carin Lundmark, Eva Schmidt, Anton Paier and Sofia Rodriguez. We also thank Åsa Bergström for technical consultation, Margaret Warner for constructive editing of the manuscript, and Adam Glick for kindly sharing his K5tTA mice. This work was supported by grants from the Tore Nilsson Foundation, the Åke Wiberg Foundation, the Hagelen Foundation, the Loo and Hans Osterman Foundation, the Torsten and Ragnar Söderberg Foundations, the Jeansson Foundations, the Swedish Research Council, and the Swedish Foundation for Strategic Research. M.H. is supported by a postdoctoral fellowship from the Wennergrenska samfundet.