Introduction

Keratins belong to the family of intermediate filament proteins and form cytoskeletal networks within all epithelial cells (Omary et al., 2004). Keratins are expressed in a tissue- and differentiation-specific manner and it is predicted that there is some degree of functional redundancy in tissues where multiple keratins are expressed (Coulombe and Omary, 2002). Of the 54 epithelial keratin genes in the human genome (Schweizer et al., 2006), 21 are currently known to be linked to human diseases characterized by fragility of specific subsets of epithelial tissues. Because keratins form highly polymeric structures, the vast majority of keratin mutations, mainly missense or small in-frame insertion/deletion mutations, disrupt cytoskeletal function by dominant-negative interference (Irvine and McLean, 1999). Gene therapy aimed at these dominant diseases might therefore specifically ablate the mutant allele; however, this requires designing allele-specific silencing reagents for each individual mutation (Kaspar, 2005; Lewin et al., 2005). Most keratin disorders are genetically heterogeneous in terms of having two or more paired keratin genes that phenocopy the disorder, each of which has an associated spectrum of pathogenic mutations (Lane and McLean, 2004). Therefore, designing specific silencing reagents for each mutation is a considerable task in terms of inhibitor design, safety testing, and passing regulatory hurdles for human use. However, in some epithelial tissues where functional redundancy of keratin genes has been shown, it may be possible to inhibit both wild-type and mutant alleles to therapeutic effect. The palmoplantar epidermis is one such tissue where at least 10 epithelial keratins are expressed (Swensson et al., 1998); therefore, this tissue, representing a relatively small area of the epidermis and one in which keratin redundancy might be exploited, is a favorable site for testing new therapies aimed at cutaneous gene inhibition (Kaspar, 2005).

Pachyonychia congenita (PC) is a rare, autosomal-dominant keratin disorder predominantly characterized by hypertrophic nail dystrophy, palmoplantar keratoderma, oral leukokeratosis, and other ectodermal defects (Leachman et al., 2005; Smith et al., 2005). Currently, there is no effective treatment for the disease. PC can be subdivided into two major variants: PC-1 (OMIM no. 167200) and PC-2 (OMIM no. 167210) as reviewed recently (Leachman et al., 2005; Smith et al., 2005). At the molecular level, heterozygous mutations in keratin K6a or K16 cause PC-1 whereas those in K6b or K17 cause PC-2 (Bowden et al., 1995; McLean et al., 1995; Smith et al., 1998). Of the epithelial tissues that are disrupted by these mutant keratins in PC patients, the main therapeutic target is the plantar epidermis, because focal keratoderma is the most painful and debilitating aspect of the disorder (Leachman et al., 2005). It has also been shown that individuals carrying mutations in K16, and presumably other PC-related keratins, can present clinically with focal plantar keratoderma alone, without other obvious ectodermal phenotypes (Shamsher et al., 1995; Smith et al., 2000, 2005). This further expands the patient population who might benefit from therapies aimed at PC, and furthermore, lessons learned from PC may well be adapted for other keratin genes.

Knockout mouse studies of the two major K6 genes, K6a and K6b, have shown that these genes exhibit functional redundancy – knocking out one of these keratins has little or no phenotypic consequences and only when both are ablated does an overt phenotype occur in the oral epithelia (Wojcik et al., 2000, 2001; Wong et al., 2000; Wong and Coulombe, 2003). Importantly, these studies showed that knockout of either K6a or K6b produces no obvious effect on the palmoplantar epidermis in mice. Of particular importance to this study, a number of recent reports have shown that the human genome exhibits far more structural variation between individuals in the population than was previously recognized, as reviewed recently (Kehrer-Sawatzki, 2007). Specifically, more than 3,500 regions of the genome have so far been proven by multiple molecular methods to exhibit copy number variation (CNV), whereby relatively large stretches of DNA, often involving entire genes or even clusters of genes, are either duplicated or deleted in individuals within the normal human population (Iafrate et al., 2004; Hinds et al., 2006; McCarroll et al., 2006; Redon et al., 2006; Wong et al., 2007). The most recent and comprehensive analysis of CNV has shown that the type II keratin locus on 12q is involved in a number of CNVs (Wong et al., 2007). Of particular interest, in relation to PC, it was shown that 3 out of 95 normal individuals analyzed carried a heterozygous CNV involving deletion of both the K6a and K6b genes and a number of nearby type II hair keratin genes (Wong et al., 2007). CNV has been shown to behave in a normal Mendelian manner and therefore from the Hardy–Weinberg calculations, about 1 in 4,000 individuals in the human population will be homozygous for this deletion and therefore be null for K6a and nearby genes (http://www.changbioscience.com/genetics/hardy.html). Because this would equate to about 75,000 individuals in the United States alone, we hypothesize that ablation of K6a in humans does not produce a PC-like phenotype or else this very large number of individuals would have come to clinical attention. From the published studies on CNV in the human genome, there is no evidence to suggest that these variants are not inherited in a normal Mendelian manner nor is there any evidence for lack of the Hardy–Weinberg equilibrium in the population (Kehrer-Sawatzki, 2007). Thus, in view of the evidence from transgenic mice and the emerging CNV data in humans, we hypothesize that specific inhibition of K6a expression, without altering the expression of K6b or the other palmoplantar keratins, may have the potential to treat the majority of PC patients, regardless of their individual mutations.

RNA interference (RNAi) is a naturally occurring gene-silencing mechanism originally identified in Caenorhabditis elegans and subsequently in mammalian cells, where it is now routinely used as a research tool (Rana, 2007). This process of sequence-specific, post-transcriptional inhibition of gene expression has great potential to be developed as a novel therapeutic approach for a number of disorders where gene inhibition is predicted to be therapeutic (Bumcrot et al., 2006). Skin disorders represent a good model for RNAi therapy development because localized cutaneous application of small interfering siRNA (siRNA) may be easier to achieve than systemic administration or use of integrating viral vector systems, both for proof-of-concept experiments in cell culture or animal models and ultimately, treatment of human subjects (Lewin et al., 2005).

Here, we present studies aimed at developing RNAi for localized treatment of plantar keratoderma in PC and, by targeting the 3'-untranslated region (3'-UTR) sequences often ignored by siRNA design algorithms, we demonstrate surprisingly potent efficacy of these gene inhibitors in both keratinocyte cultures and a mouse model.

Results

Inhibition of K6a expression

DNA sequence alignments of the cDNAs encoding human K6a (Genbank RefSeq accession no. NM_005554) and K6b (Genbank RefSeq accession no. NM_005555) revealed that only a few isolated bases can distinguish these two genes in terms of their protein-encoding sequences (data not shown). However, these two genes differ significantly in their non-coding 3'-UTR sequences, allowing the design of four inhibitors (designated K6a-S1, -S2, -S3, and -S4; see Table 1), which were predicted to inhibit K6a expression without affecting the expression of K6b or other type II keratin genes due to significant sequence differences (Figure 1a). The four K6a inhibitors were initially tested in human 293FT embryonic kidney cells for inhibition of enhanced green fluorescent protein (EGFP)-tagged K6a. Either K6a-specific, EGFP-specific, or non-specific control siRNAs were co-transfected into 293FT cells along with a K6a-enhanced yellow fluorescent protein (EYFP) expression plasmid. IC₅₀ is a unit of measurement for the efficacy of inhibitors, such as siRNA, and in this case is defined as the inhibitor concentration resulting in a 50% reduction in gene expression. All four K6a-specific siRNAs strongly inhibited K6a expression, giving highly favorable IC₅₀ values between 0.03 and 0.1 nM as determined by fluorescence-activated cell sorting analysis of cells expressing the EYFP-tagged K6a protein (Figure 1b). As a positive control, EGFP-specific siRNA inhibitor co-transfected with the K6a-EYFP plasmid gave an IC₅₀ of 0.1 nM (Figure 1b). The sequence target of this EGFP control inhibitor is conserved in EYFP and other modified fluorescent proteins derived from EGFP and it is known to be a potent siRNA in cultured cells and in vivo using a mouse model system (Wang et al., 2007). Thus, our K6a siRNAs were comparable in potency to a recognized potent inhibitor, especially at the low, picomolar concentrations that would be useful for therapeutic application (Figure 1b). In contrast, no effect was observed with the irrelevant control (NSC4) siRNA inhibitor (Figure 1b). Cells were also visualized by fluorescence microscopy (Olympus CK40) 48 hours after transfection using an EGFP filter set (Figure 1c), confirming by a different, more visual method that these inhibitors are effective at picomolar concentrations. No changes in cell density or morphology were observed by bright field imaging (data not shown).

Figure 1.

Sequence specificity and efficacy of K6a 3'-UTR siRNAs. (a) Alignments of portions of the human K6a and K6b 3'-UTR cDNA sequences, showing that the target sequences of the four siRNAs (bold type) are located where several insertion/deletion and single-base substitutions occur between K6a and K6b. (b) Normalized K6a-EYFP expression measured by FACS of 293FT cells transiently transfected with K6a-EYFP and the lead K6a siRNAs (K6a-S1, -S2, -S3, and -S4). A potent EGFP siRNA (which also targets EYFP equally well) was used as a positive control. A non-specific control siRNA (NSC4) did not significantly affect expression, even at high concentrations, whereas all four K6a inhibitors worked as well or better than the EGFP control inhibitor at low sub-nanomolar concentrations. Bars on the graph represent the meanSE. (c) Visual confirmation of K6a siRNA efficacy by wide-field imaging of 293FT cells transiently transfected with K6a-EYFP and siRNAs over a range of inhibitor concentrations.

Full figure and legend (239K)

Table 1 - siRNA sequences.

Full table

Inhibition of endogenous K6a in HaCaT and NEB-1 cells

Currently, there are no antibodies available to distinguish between human K6a and K6b proteins and it is unlikely that such reagents will emerge because there are only a few isolated amino acids that differ between these proteins. However, reverse transcriptase-PCR analysis using primers specific to either K6a or K6b demonstrated that both immortalized keratinocyte cell lines HaCaT and NEB-1 cells express significant amounts of K6a and negligible amounts of K6b (data not shown). This is in comparison with primary keratinocyte cultures, which we have found to express high levels of both K6a and K6b by reverse transcriptase-PCR (data not shown). HaCaT and NEB-1 are therefore suitable cell lines to test the efficacy of K6a siRNAs because inhibition of K6a will not be masked by the expression of K6b. Initially, the four K6a siRNAs were independently transfected at a final concentration of 5 nM into HaCaT cells. Total protein extracts were prepared at 24, 48, 72, 96, and 120 hours post-transfection. Immunoblotting with an anti-K6 antibody showed approximately 50% reduction of K6 protein expression with each of the four K6a siRNAs by 48 hours, which increased to almost 100% loss of expression by 72–96 hours (data not shown). From a separate ongoing study in the laboratory, the half-life of epidermal keratins is of the order of 12–24 hours, consistent with the clearance time observed here (McLean, unpublished data). Transfections were repeated and cytoskeletal extracts were prepared 96 hours after transfection for protein staining of SDS–PAGE gels and immunoblot analysis with a range of antibodies (Figure 2). A highly significant decrease in the levels of K6 was visualized in protein gels stained with Coomassie blue; all other cytoskeletal proteins were unaffected (Figure 2a). This was confirmed by staining blots with antibodies against K6, K5, K8, K14, K16, K17, K18, K19, and lamin A/C. Although there was almost 100% reduction of K6 with each of the four K6a siRNAs, little or no change was observed in the levels of all other keratins analyzed (Figure 2b). Interestingly, transfection with K6a-S3 did however lead to a slight reduction in the amount of lamin A/C, indicative of possible off-target effects (Figure 2b). The target sequence of K6a-S3 is not well conserved in lamin A/C (which is also an intermediate filament protein distantly related to keratins); therefore, this result, which was consistent in replicate experiments in both HaCaT and NEB-1 cells, does not have an obvious explanation. The K6a-S3 sequence is, however, somewhat repetitive and this may explain the non-specific effects. Thus, K6a-S3 is not of sufficient specificity for further development toward human use. All these efficacy and specificity experiments were repeated on the NEB-1 cell line (Morley et al., 2003) with identical results obtained (data not shown).

Figure 2.

Inhibition of endogenous K6a in HaCaT cells. (a) Coomassie blue staining of cytoskeletal extracts 96 hours after transfection with 5 nM K6a siRNAs showing near-complete loss of K6a expression with all four K6a 3'-UTR siRNAs; all other keratins were unaffected. (b) Immunoblotting of replicate gels as shown in (a), with a range of anti-keratin antibodies, confirming the loss of K6a expression and normal expression of other keratins present in HaCaT cells.

Full figure and legend (159K)

Specificity of K6a siRNAs

To confirm that the siRNAs did not affect the expression of K6b, the most obvious candidate for off-target effects, a dual-expression construct was made that expresses both FLAG-tagged K6a and myc-tagged K6b to enable detection of each K6 protein using antibodies against the specific epitope tags. This construct was co-transfected into AD293 cells with each of the four K6a siRNAs and total protein extracted 48 hours after transfection. Analysis of extracts by Western blotting showed essentially complete loss of K6a protein, whereas levels of K6b were only slightly reduced (Figure 3).

Figure 3.

K6a 3'-UTR siRNAs do not significantly affect K6b expression. (a) Immunoblot of total protein extracts from AD293 cells transfected with dual expression vector expressing FLAG epitope-tagged K6a and myc epitope-tagged K6b in combination with test siRNAs. Blot stained with Ponceau S to demonstrate equal protein loading in each lane. (b) Replica immunoblot shown in (a), stained with anti-FLAG antibody to show K6a-FLAG expression. All four K6a siRNAs completely knock out K6a expression, whereas cells transfected with K6a/K6b plasmid alone or an irrelevant siRNA show high levels of K6a-FLAG expression. (c) Replica immunoblot shown in (a), stained with anti-myc antibody to show K6b-myc expression. Only a slight reduction in K6b-myc expression is seen with the four K6a-specific siRNAs.

Full figure and legend (100K)

K6a-S1 strongly inhibits reporter gene expression in vivo

Previously, we have shown that specific siRNAs (but not non-specific controls) strongly inhibit epidermal EGFP/luciferase transgene expression in vivo by intradermal injection into mouse footpads (Wang et al., 2007). One of the potentially therapeutic siRNAs developed here, K6a-S1, was chosen for in vivo testing in mouse footpad by co-injection with a bicistronic K6a-luciferase reporter construct. Using luciferase activity as an in vivo readout, we were able to show that K6a-S1, but not nonspecific controls, strongly inhibited K6a/luciferase gene expression in mouse skin (Figure 4). Averaging the bioluminescence readout from several mice, K6a/luciferase reporter gene expression was found to be strongly inhibited at the earliest time point examined (1 day) and remained barely detectable over the 3-day course of the experiment (Figure 4).

Figure 4.

K6a-S1 is effective in a mouse footpad in vivo assay. (a) Mouse footpads were injected with a bicistronic K6a-luciferase reporter gene plasmid in both footpads. The left footpad in each mouse was co-injected with an inert pUC18 plasmid in place of siRNA as a positive control; the right footpad was co-injected with K6a-S1 siRNA. The amount of nucleic acid injected in each footpad was standardized using pUC18 DNA. (b) Owing to the inherent variability in these experiments (for example, the positive control footpad in mouse 4 did not express the reporter gene at all), several replicates were carried out and the bioluminescence signals captured with the Xenogen IVIS system were averaged and plotted. Five mice from the day 2 time point are shown. (b) Quantitation of K6a-luciferase gene expression shows that on average, K6a/luciferase expression is greatly reduced by siRNA K6a-S1. The wide error bars in the control footpads reflect the inherent variability of these experiments; however, the normalized siRNA data show consistent and highly potent knockdown of the K6a reporter gene.

Full figure and legend (81K)

Discussion

Treatment of autosomal-dominant skin disorders such as PC requires the development of effective methods to reduce or preferably, fully inactivate, the dominant-negative mutant allele (Kaspar, 2005; Lewin et al., 2005). The recently identified RNAi pathway and in particular, the availability of siRNA specific to user-defined genetic sequence targets offers a novel way to achieve this (Bumcrot et al., 2006; Rana, 2007). In the case of PC, an unusual situation exists where there is more than one copy of the affected gene (K6a and K6b), with similar, overlapping expression patterns (Smith et al., 1998) and where complete loss of K6a is predicted from mouse knockouts to be tolerated owing to the compensatory expression of K6b (Wojcik et al., 2000, 2001; Wong et al., 2000; Wong and Coulombe, 2003). Furthermore, the recently identified CNV in the human genome strongly suggests that homozygous ablation of the K6a/K6b genes exists in as many as 1 in 4,000 humans, without obvious pathology (Wong et al., 2007). This therefore opens up the possibility for a gene-silencing approach that is independent of the particular K6a mutation carried by an individual patient. Careful genetic analysis of many affected families has shown that K6a is the most common gene mutated in PC (Smith et al., 2005). Therefore, this approach, if successful, could treat the majority of patients with this disease using a single gene-silencing reagent. This has the considerable added advantage that only one reagent would require toxicity testing, manufacture to current Good Manufacturing Practice standards, and approval for human use from the Food and Drug Administration or other equivalent regulatory agencies outside the United States, all of which represent considerable obstacles to therapy development for orphan diseases affecting small numbers of individuals (Epstein et al., 2006).

By targeting the 3'-UTR of the K6a gene, we were able to identify four highly efficacious siRNAs capable of essentially ablating K6a expression in keratinocyte cell lines by 72–96 hours. The general guidelines for siRNA design suggest that the 5'- and 3'-UTR sequences should be avoided as the target site (Reynolds et al., 2004). The rationale is that some genes may have alternate spliced exons that generate mRNA species with alternate 5'- and/or 3'-UTRs, so that there is a risk of not fully ablating all transcripts of the gene of interest. In the case of the keratins, these genes and their transcripts have been well studied and no alternate splicing has been reported. Furthermore, the human genome project has made available many expressed sequence tags that align to the K6a and K6b genes and none of these show alternate splicing (http://genome.ucsc.edu). Thus, with keratin genes at least, the 3'-UTR sequences represent a good target for siRNA. In fact, the potency of our K6a 3'-UTR inhibitors was so great that even Coomassie blue staining readily showed their powerful inhibitory effects (Figure 2a). When one considers how abundant keratin mRNAs and proteins are within the cytoplasm of keratinocytes and keratinocyte cell lines (Lane, 1993), it is remarkable to see near-ablation at the protein level with picomolar amounts of siRNA.

The coding sequences of keratins show a high degree of sequence conservation and therefore exploiting the K6a 3'-UTR sequence, which is similar to intronic or intergenic sequences in terms of its poor intergene and interspecies conservation, has the further advantage of avoiding potential off-target effects against any of the other 52 functional keratin genes in the human genome in addition to its closest relative, K6b. Only one of the four siRNAs identified here, K6a-S3, showed any off-target effects against the more distantly related lamin A/C, which may be due to the presence of repeat sequences within this particular target. This leaves three lead inhibitors for further development and of these, one has already been shown to be effective in an in vivo mouse footpad model (Figure 4).

From studies of animal models of keratin disorders, it is predicted that perhaps even partial reduction of a mutant keratin allele might lead to significant improvement in the clinical phenotype. Specifically, studies of mice with inducible dominant-negative mutations in K14 (an inducible mouse model of epidermolysis bullosa simplex), have demonstrated in mice that even 50% knockdown of the mutant allele would prevent the development of epidermal blistering (Cao et al., 2001). There remains the issue of the structural and physiological differences between mouse and human epidermis, particularly in terms of the levels of mechanical stress experienced in everyday life; however, these data strongly emphasize the potential of even partial gene inhibition as a route to therapy for many human keratin disorders, including PC.

As we have demonstrated here, siRNA technology has great potential as a therapeutic approach for treatment of dominant-negative skin diseases. The reagents themselves fall somewhere between small molecules and gene therapy in terms of their size and chemical characteristics and so the major hurdle now faced in the development of siRNA for therapeutic use is the problem of delivery through the stratum corneum into the living epidermis. Despite the advantage of skin as one of the most accessible organs in the body, one of its primary functions is to provide a physical barrier to prevent the entry of foreign pathogens, allergens, and irritants. As highlighted in a recent conference/workshop to discuss obstacles to research translation with particular emphasis on dermatology (Epstein et al., 2006), delivery into skin is an area that requires further work if siRNA therapy for PC or other skin diseases is to become a reality. A further problem with keratinizing disorders such as PC is that the very thick keratoderma would potentially act as a further barrier to impede the entry of the inhibitor into the living cells of the epidermis where the gene-silencing effect is required, at least in the early stages of treatment.

In conclusion, we show that targeting the 3'-UTR of the K6a gene allows both potent and specific inhibition of K6a expression both in cultured cells and early in vivo studies. Coupled with a suitable epidermal delivery system, these reagents could provide a potential route to therapy for PC, an incurable and highly debilitating genetic skin disorder.

Materials and Methods

siRNA design

Four siRNAs (designated K6a-S1, -S2, -S3, and -S4) were designed (19+2 format) against the unique 3'-UTR region of K6a (Table 1) and synthesized by Dharmacon Inc., (Layfayette, CO). Lamin A/C and EGFP-specific siRNAs were used as controls (Dharmacon).

Expression constructs

A wild-type K6a fusion protein tagged with EYFP was generated as described previously (Hickerson et al., 2006). To make a dual expression plasmid for expression of epitope-tagged versions of K6a and K6b, a full-length human K6a IMAGE clone 3639270 was subjected to site-directed mutagenesis using the Stratgene QuikChange system. Specifically, the termination codon was ablated and replaced by a unique RsrII restriction site, allowing in-frame insertion of an oligo cassette, with RsrII cohesive ends, encoding the FLAG epitope DYKDDDDK, followed by an in-frame stop. By the same strategy, the termination codon of a full-length K6b IMAGE clone 4754558 was replaced by an RsrII site allowing insertion of an oligo cassette encoding the myc epitope EQKLISEEDL. The FLAG-tagged K6a construct was blunt-cloned into the XhoI site of the pBudCE4.1dual expression plasmid (Invitrogen, Paisley, UK) for expression under the EF-1a promoter; the myc-tagged K6b construct was blunt-cloned into the XbaI site of pBudCE4.1 for expression under the cytomegalovirus promoter. A bicistronic construct expressing both firefly luciferase (fLUC) and wild-type K6a was prepared as described previously (Hickerson et al., 2006).

Transient transfections

Human 293FT embryonic kidney cells (Invitrogen) were maintained in DMEM (Sigma Chemicals, Poole, UK) supplemented with 10% fetal bovine serum (Invitrogen). Cells were seeded onto 48-well plates at 8 10⁴ cells/well on the day before transfection. At approximately 80% confluence, they were transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Cells were co-transfected with K6a-EYFP (150 ng) and the specific K6a 3'-UTR siRNAs (K6a-S1, -S2, -S3, or -S4) or a non-specific control (NSC4) siRNA (final concentration of 0.016–4 nM) supplemented with pUC19 plasmid DNA to give a final nucleic acid concentration of 800 ng per transfection. Cells were visualized by fluorescence microscopy (Olympus CK40) 48 hours following transfection using an EGFP filter set. To quantitate siRNA inhibition, cells were trypsinized 48 hours after transfection and analyzed by FACS (Becton Dickinson FACScan) using channel FL1 (530 nm emission filter). Ten thousand cells per transfection were analyzed. The data were generated by gating the cells and determining the percentage of cells that decreased below the gate with and without siRNA treatment. The data were normalized to 100 and then corrected against data from cells treated with NSC4 siRNA.

HaCaT (Boukamp et al., 1988) and NEB-1 cells (Morley et al., 2003) were routinely grown in DMEM plus 10% fetal bovine serum. NEB-1 cells were grown in RM+ medium, which consisted of DMEM containing 25% Ham's F12 medium (Sigma), 10% fetal bovine serum, plus additional growth factors (hydrocortisone (0.4 g/ml), cholera toxin (10^-10 M), transferrin (5 g/ml), lyothyronine (2 10^-11 M), adenine (1.9 10^-4 M), and insulin (5 g/ml)). All cell lines were fibroblast feeder cell-independent and were cultured at 37°C in 5% CO₂.

HaCaT and NEB-1 cells were transfected using Lipofectamine RNAiMax (Invitrogen) according to the manufacturer's protocol for "reverse transfection." Each K6a siRNA or lamin A/C control siRNA, was diluted in 500 l Opti-MEM medium (Invitrogen). Lipofectamine RNAiMax (7 l) was added, mixed, and incubated at room temperature for 20 minutes. Trypsinized HaCaT or NEB-1 cells (2.5 10⁵ cells in 2.5 ml growth medium) were added to each 3 cm² plate. The final siRNA concentration was 5 nM. Cells were incubated at 37°C for 24, 48, 72, 96, and 120 hours.

Human AD293 embryonic kidney cells (Graham et al., 1977) were maintained in DMEM plus 10% fetal bovine serum. Cells were seeded onto 3 cm² plates and transfected at 80–90% confluence using Lipofectamine 2000 according to manufacturer's instructions. Briefly 500 ng plasmid pBudCE4.1 vector (Invitrogen), co-expressing K6a-FLAG and K6b-c-Myc were diluted in 250 l Opti-MEM medium and K6a 3'-UTR siRNAs or lamin A/C siRNA added. Lipofectamine 2000 (3 l) was diluted in 250 l Opti-MEM medium. Diluted Lipofectamine 2000 and diluted DNA/siRNA were mixed together, incubated at room temperature for 20 minutes and added to the cells. The final siRNA concentration was 0.25 nM. Cells were incubated at 37°C for 48 hours.

Protein biochemistry

Cultured cells were washed twice with phosphate-buffered saline and scraped from wells. For total protein extracts, cells were pelleted at 1,000 r.p.m. for 5 minutes, and resuspended in 400 l NuPAGE lithium dodecyl sulfate sample buffer (Invitrogen).

For cytoskeletal extracts, cultured cells were washed twice with phosphate-buffered saline, followed by incubation on ice in low salt buffer (10 mM Tris–HCl, 150 mM NaCl, 3 mM EDTA, 0.1% NP-40 (pH 7.4) with HCl) for 20 minutes. The buffer was replaced with high salt buffer (10 mM Tris–HCl, 150 mM NaCl, 1.5 M KCl, 3 mM EDTA, 0.1% NP-40, pH 7.4) and incubated for a further 20 minutes on ice. Cells were scraped off and pelleted at 3,500 r.p.m. for 15 minutes. Pellets were washed three times in wash buffer (10 mM Tris–HCl, 150 mM NaCl, 3 mM EDTA, pH 7.4) and centrifuged for 15 minutes at 3,500 r.p.m. between washes. Final pellets were resuspended in 200 l NuPAGE lithium dodecyl sulfate sample buffer (Invitrogen). All extraction and washing buffers contained protease inhibitor cocktail (0.1%, Sigma).

Cytoskeletal/total protein samples were heated to 70°C for 10 minutes in the presence of reducing agent (final concentration of 50 mM dithiothreitol; Invitrogen), and resolved on 4–12% NuPAGE Bis/Tris gels (Invitrogen) with SDS-MOPS running buffer. SeeBlue Prestained Standards (5 l; Invitrogen) were run alongside as protein size markers. Gels were either stained with SimplyBlue SafeStain (Invitrogen) or transferred to nitrocellulose membrane for immunostaining. Efficiency of transfer and equal loading of samples was confirmed by staining with Ponceau S solution (Sigma). Membranes were blocked in TBS (10 mM Tris–HCl (pH 8.0), 0.15 M NaCl) containing 5% marvel and/or 0.05% Tween-20 (Sigma) for between 1–4 hours.

Immunoblotting

Keratin expression was detected by immunoblotting using 1 hour incubations of the following primary antibodies: 1:200 dilution anti K6 (Ks6.KA12; Progen Biotechnik, GmbH, Heidelberg, Germany); 1:500 dilution anti K17 (CK-E3, Sigma); anti K16 (LL025); anti K5 and K8 (RCK102); 1:1,000 anti K14 (LL001); 1:10 anti K19 (LP2K); 1:2 dilution anti K18 (LDK18); 1:10,000 dilution anti K5 (BL18)(all kindly donated by E.B. Lane, College of Life Sciences, University of Dundee, Dundee, UK); 1:200 dilution lamin A/C (Santa Cruz Biotechnology, Santa Cruz, CA); 1:1,000 dilution anti-FLAG tag rabbit polyclonal (anti-DDDDK tag.ab1162; Abcam Ltd, Cambridge, UK); and 1:800 dilution 9E10 anti c-Myc (BD Biosciences Pharmigen, Cowley, UK). After primary antibody incubation, blots were washed three times for 5 minutes in 0.05% Tween-20/TBS and incubated in secondary antibodies (1:1,000 dilution of either alkaline phosphatase-conjugated rabbit anti-mouse immunoglobulins (D314, DakoCytomation, Ely, UK) or goat anti-rabbit immunoglobulins (D487) for 1 hour. Blots were washed as above and developed using nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl-phosphate color substrate (Promega, Southampton, UK).

Mouse footpad injections and in vivo imaging

Footpad injections into 6–8-week female FVB mice were performed as described previously (Wang et al., 2007). Phosphate-buffered saline (50 l) containing 10 g of expression plasmid and 15 g siRNA inhibitor (or pUC18 plasmid DNA) was intradermally injected (28G needle). At the indicated time points, the mice were imaged (10 minutes after intraperitoneal injection of 100 l of 30 mg/ml (150 mg/kg body weight) luciferin (Xenogen Biosciences, Cranbury, NJ)). Mice were anesthesized using isoflurane and imaged using the IVIS200 system (Xenogen Biosciences). Emitted light was quantitated with LivingImage software (Xenogen). The raw data (photons/second/cm²/srad) from footpads treated with K6a-S1 or EGFP siRNAs were corrected by dividing by the amount of light emitted from control footpads receiving pUCC18 in place of siRNAs.

Conflict of Interest

The authors state no conflict of interest.

References

1. Boukamp P, Petrussevska RT, Breitkreutz D, Hornung J, Markham A, Fusenig NE (1988) Normal keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line. J Cell Biol 106:761–771 | Article | PubMed | ISI | ChemPort |
2. Bowden PE, Haley JL, Kansky A, Rothnagel JA, Jones DO, Turner RJ (1995) Mutation of a type II keratin gene (K6a) in pachyonychia congenita. Nat Genet 10:363–365 | Article | PubMed | ISI | ChemPort |
3. Bumcrot D, Manoharan M, Koteliansky V, Sah DW (2006) RNAi therapeutics: a potential new class of pharmaceutical drugs. Nat Chem Biol 2:711–719 | Article | PubMed | ISI | ChemPort |
4. Cao T, Longley MA, Wang X-J, Roop DR (2001) An inducible mouse model for epidermolysis bullosa simplex: implications for gene therapy. J Cell Biol 152:651–656 | Article | PubMed | ISI | ChemPort |
5. Coulombe PA, Omary MB (2002) 'Hard' and 'soft' principles defining the structure, function and regulation of keratin intermediate filaments. Curr Opin Cell Biol 14:110–122 | Article | PubMed | ISI | ChemPort |
6. Epstein EH Jr, Gilchrest BA, Milstone LM (2006) Obstacles to translation conference. J Invest Dermatol 126:1434–1437 | Article | PubMed | ISI | ChemPort |
7. Graham FL, Smiley J, Russell WC, Nairn R (1977) Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J Gen Virol 36:59–74 | PubMed | ISI | ChemPort |
8. Hickerson RP, Smith FJD, McLean WHI, Landthaler M, Leube RE, Kaspar RL (2006) SiRNA-mediated selective inhibition of mutant keratin mRNAs responsible for the skin disorder pachyonychia congenita. Ann N Y Acad Sci 1082:56–61 | Article | PubMed | ISI | ChemPort |
9. Hinds DA, Kloek AP, Jen M, Chen X, Frazer KA (2006) Common deletions and SNPs are in linkage disequilibrium in the human genome. Nat Genet 38:82–85 | Article | PubMed | ISI | ChemPort |
10. Iafrate AJ, Feuk L, Rivera MN, Listewnik ML, Donahoe PK, Qi Y et al. (2004) Detection of large-scale variation in the human genome. Nat Genet 36:949–951 | Article | PubMed | ISI | ChemPort |
11. Irvine AD, McLean WHI (1999) Human keratin diseases: the increasing spectrum of disease and subtlety of the phenotype-genotype correlation. Br J Dermatol 140:815–828 | Article | PubMed | ISI | ChemPort |
12. Kaspar RL (2005) Challenges in developing therapies for rare diseases including pachyonychia congenita. J Investig Dermatol Symp Proc 10:62–66 | Article | PubMed | ISI |
13. Kehrer-Sawatzki H (2007) What a difference copy number variation makes. Bioessays 29:311–313 | Article | PubMed | ISI | ChemPort |
14. Lane EB (1993) Keratins. In: Royce PM, Steinmann B (eds) Connective Tissue and its Heritable Disorders Molecular, Genetic and Medical Aspects. New York: Wiley-Liss Inc., 237–247
15. Lane EB, McLean WHI (2004) Keratins and skin disorders. J Pathol 204:355–366 | Article | PubMed | ISI | ChemPort |
16. Leachman SA, Kaspar RL, Fleckman P, Florell SR, Smith FJD, McLean WHI, et al. (2005) Clinical and pathological features of pachyonychia congenita. J Investig Dermatol Symp Proc 10:3–17 | Article | PubMed | ISI | ChemPort |
17. Lewin AS, Glazer PM, Milstone LM (2005) Gene therapy for autosomal dominant disorders of keratin. J Investig Dermatol Symp Proc 10:47–61 | Article | PubMed | ISI | ChemPort |
18. McCarroll SA, Hadnott TN, Perry GH, Sabeti PC, Zody MC, Barrett JC et al. (2006) Common deletion polymorphisms in the human genome. Nat Genet 38:86–92 | Article | PubMed | ISI | ChemPort |
19. McLean WHI, Rugg EL, Lunny DP, Morley SM, Lane EB, Swensson O et al. (1995) Keratin 16 and keratin 17 mutations cause pachyonychia congenita. Nat Genet 9:273–278 | Article | PubMed | ISI | ChemPort |
20. Morley SM, D'Alessandro M, Sexton C, Rugg EL, Navsaria H, Shemanko CS et al. (2003) Generation and characterization of epidermolysis bullosa simplex cell lines: scratch assays show faster migration with disruptive keratin mutations. Br J Dermatol 149:46–58 | Article | PubMed | ISI | ChemPort |
21. Omary MB, Coulombe PA, McLean WHI (2004) Intermediate filament proteins and their associated diseases. N Engl J Med 351:2087–2100 | Article | PubMed | ISI | ChemPort |
22. Rana TM (2007) Illuminating the silence: understanding the structure and function of small RNAs. Nat Rev Mol Cell Biol 8:23–36 | Article | PubMed | ISI | ChemPort |
23. Redon R, Ishikawa S, Fitch KR, Feuk L, Perry GH, Andrews TD et al. (2006) Global variation in copy number in the human genome. Nature 444:444–454 | Article | PubMed | ISI | ChemPort |
24. Reynolds A, Leake D, Boese Q, Scaringe S, Marshall WS, Khvorova A (2004) Rational siRNA design for RNA interference. Nat Biotechnol 22:326–330 | Article | PubMed | ISI | ChemPort |
25. Schweizer J, Bowden PE, Coulombe PA, Langbein L, Lane EB, Magin TM et al. (2006) New consensus nomenclature for mammalian keratins. J Cell Biol 174:169–174 | Article | PubMed | ISI | ChemPort |
26. Shamsher MK, Navsaria HA, Stevens HP, Ratnavel RC, Purkis PE, Kelsell DP et al. (1995) Novel mutations in keratin 16 gene underly focal non-epidermolytic palmoplantar keratoderma (NEPPK) in 2 families. Hum Mol Genet 4:1875–1881 | Article | PubMed | ISI | ChemPort |
27. Smith FJD, Fisher MP, Healy E, Rees JL, Bonifas JM, Epstein EH Jr et al. (2000) Novel keratin 16 mutations and protein expression studies in pachyonychia congenita type 1 and focal palmoplantar keratoderma. Exp Dermatol 9:170–177 | Article | PubMed | ISI | ChemPort |
28. Smith FJD, Jonkman MF, van Goor H, Coleman C, Covello SP, Uitto J et al. (1998) A mutation in human keratin K6b produces a phenocopy of the K17 disorder pachyonychia congenita type 2. Hum Mol Genet 7:1143–1148 | Article | PubMed | ISI | ChemPort |
29. Smith FJD, Liao H, Cassidy AJ, Stewart A, Hamill KJ, Wood P et al. (2005) The genetic basis of pachyonychia congenita. J Investig Dermatol Symp Proc 10:21–30 | Article | PubMed | ISI | ChemPort |
30. Swensson O, Langbein L, McMillan JR, Stevens HP, Leigh IM, McLean WHI et al. (1998) Specialized keratin expression pattern in human ridged skin as an adaptation to high physical stress. Br J Dermatol 139:767–775 | Article | PubMed | ISI | ChemPort |
31. Wang Q, Ilves H, Chu P, Contag CH, Leake D, Johnston BH et al. (2007) Delivery and Inhibition of Reporter Genes by Small Interfering RNAs in a Mouse Skin Model. J Invest Dermatol [e-pub ahead of print 24 May 2007, doi:10.1038/sj.jid.5700891] | Article | PubMed | ChemPort |
32. Wojcik SM, Bundman DS, Roop DR (2000) Delayed wound healing in keratin 6a knockout mice. Mol Cell Biol 20:5248–5255 | Article | PubMed | ISI | ChemPort |
33. Wojcik SM, Longley MA, Roop DR (2001) Discovery of a novel murine keratin 6 (K6) isoform explains the absence of hair and nail defects in mice deficient for K6a and K6b. J Cell Biol 154:619–630 | Article | PubMed | ISI | ChemPort |
34. Wong KK, deLeeuw RJ, Dosanjh NS, Kimm LR, Cheng Z, Horsman DE et al. (2007) A comprehensive analysis of common copy-number variations in the human genome. Am J Hum Genet 80:91–104 | Article | PubMed | ISI | ChemPort |
35. Wong P, Colucci-Guyon E, Takahashi K, Gu C, Babinet C, Coulombe PA (2000) Introducing a null mutation in the mouse K6alpha and K6beta genes reveals their essential structural role in the oral mucosa. J Cell Biol 150:921–928 | Article | PubMed | ISI | ChemPort |
36. Wong P, Coulombe PA (2003) Loss of keratin 6 (K6) proteins reveals a function for intermediate filaments during wound repair. J Cell Biol 163:327–337 | Article | PubMed | ISI | ChemPort |

Acknowledgments

FJDS is supported by a Career Development Fellowship from The Pachyonychia Congenita Project (http://www.pachyonychia.org). Research in the Smith/McLean laboratory is also supported by grants from The Dystrophic Epidermolysis Bullosa Research Association UK and the British Skin Foundation/National Eczema Society.