Modulation of Mouse Coagulation Gene Transcription following Acute In Vivo Delivery of Synthetic Small Interfering RNAs Targeting HNF4α and C/EBPα

Hepatocyte nuclear factor 4α (HNF4α) and CCAAT/enhancer-binding protein α (C/EBPα) are important for the transcriptional control of coagulation factors. To determine in vivo the direct role of HNF4α and C/EBPα in control of genes encoding coagulation factors, a synthetic small interfering (si)RNA approach was used that enabled strong reduction of mouse hepatic HNF4α and C/EBPα under conditions that minimized target-related secondary effects. For both HNF4α and C/EBPα, intravenous injection of specific synthetic siRNAs (siHNF4α and siC/EBPα) resulted in more than 75% reduction in their liver transcript and protein levels 2 days post-injection. For siHNF4α, this coincided with marked and significantly reduced transcript levels of the coagulation genes Hrg, Proz, Serpina5, F11, F12, F13b, Serpinf2, F5, and F9 (in order of magnitude of effect) as compared to levels in control siRNA injected animals. Significant decreases in HNF4α target gene mRNA levels were also observed at 5 days post-siRNA injection, despite a limited level of HNF4α knockdown at this time point. Compared to HNF4α, C/EBPα knockdown had a modest impact on genes encoding coagulation factors. A strong reduction in C/EBPα transcript and protein levels resulted in significantly affected transcript levels of the control genes Pck1 and Fasn and a modest downregulation for coagulation genes Fba, Fbg and F5. F5 and F11 were the sole coagulation genes that were significantly affected upon prolonged (5 day) C/EBPα knockdown. We conclude that in the mouse, HNF4α has a direct and essential regulatory role for multiple hepatic coagulation genes, while a role for C/EBPα is more restricted. In addition, this study demonstrates that synthetic siRNA provides a simple and fast means for determining liver transcription factor involvement in vivo.

The in vivo importance of HNF4a in regulating hepatic transcription of coagulation genes described in studies using hepatocyte-specific HNF4a knockout mice [14,15]. Hnf4a disruption affected expression of factor (F) F5, F9, F11, F12, F13b, protein C inhibitor (Serpina5), protein Z (Proz), a2-antiplasmin (Serpinf2), protein Z inhibitor (Serpina10), and heparin Cofactor II (Serpind1), whereas no effects were observed for F2, F7, F8, F10, protein S (Pros1) and antithrombin (Serpinc1) [14,15]. For C/EBPa, studies on the relevance to coagulation are limited. A recent CHIPseq study determined the genome-wide occupancy of C/ EBPa in the livers of human, mouse, dog, opossum, and chicken. Fibrinogen A (FGA) and F2 were identified among 32 genes located near 35 C/EBPa binding events that were conserved among these five vertebrates [16]. In the mouse genome, Fga and F2, C/EBPa binding sites were 45 and 64 bp from the transcription start site, respectively. However, it should be emphasized that binding does not necessarily correlate with functional activity in controlling gene transcription. Furthermore, carriers of hemophilia B Leiden have a causal mutation in a C/ EBPa binding site in F9 promoter [17], and in line, liver-specific C/EBPa-null mice display reduced hepatic expression of F9 [17,18]. Finally, in vitro studies demonstrated requirement of C/ EBPa for F8 expression, but this involved nonhepatic inflamed cells [19]. Whether other coagulation genes other than FGA, F2, F9 and F8 are regulated by C/EBPa is unknown.
The study of the in vivo roles of HNF4a and C/EBPa in control of gene transcription in liver employed a conditional gene knockout approach because conventional gene knockouts for HNF4a and C/EBPa were embryonic lethal [20,21]. In general, a conditional gene knockout approach does not allow a rapid significant deletion of the gene of interest in vivo without challenging liver physiology; fast adenovirus-mediated hepatic delivery of the required Cre-recombinase does allow rapid hepatic disruption of 'floxed' alleles but is concomitant with adenovirusrelated acute hepatic inflammation [22,23]. Similarly, inducible liver-specific gene disruption based on the MX1-Cre transgene requires a burst of circulating interferon to evoke the necessary activation of MX1-Cre [24]. Meaningful studies of the role of hepatic transcription factors can therefore only start after weaning of the adenovirus or interferon effects. At that time, transcription factor deletion may already have induced (secondary) changes in liver physiology and compensatory changes in expression of other hepatic transcription factors. It may lead to a possible misinterpretation of the direct role of a given transcription factor in gene regulation. Hence, the current observations regarding the role of HNF4a and C/EBPa in regulating coagulation gene transcription obtained in adult mice lacking HNF4a and C/EBPa from birth on (Cre recombinase under control of the albumin promoter) [25], following Cre supplied by means of adenovirus [26] or the MX1-Cre transgene [27] may be in part secondary to changes in liver physiology and changes in expression of other hepatic transcription factors. This may, explain the unexpected transcriptional increase of numerous hepatically expressed genes including the coagulation gene Serpind1 and absence of effects for the in vitro identified targets F2, F7, F8, F10, Pros1 and Serpinc1 in livers from mice lacking HNF4a from birth [14,15].
Recently, lipid-based reagents became available that allow efficient delivery of synthetic small interfering (si) RNAs to livers of adult mice following systemic injection [28]. Thus, transient knockdown of target gene expression can be achieved rapidly (within two days post siRNA delivery) and does not involve changes in liver physiology as a result of harsh methodology. In the present study, we used this in vivo siRNA approach to rapidly reduce HNF4a and C/EBPa expression in mouse livers and to determine the impact of these two distinct transcription factors on hepatic coagulation gene transcription.

siRNA screening and validation
Pre-designed siRNAs for mouse HNF4a and C/EBPa mRNAs were purchased from Ambion Applied Biosystems, Carlsbad, California, USA (Ambion SilencerH Pre-designed for HNF4a; catalogue numbers 67633 (#1), 67634 (#2) and 67635 (#3) with sense sequences GGC AGA UGA UCG AAC AGA UUU, CCA AUG UCA UUG UUG CUA AUU, and AGA GGU CCA UGG UGU UUA AUU, respectively and for mouse C/EBPa; catalogue numbers 63853 (#1), 63854 (#2), 63855 (#3) with sense sequences GCA AAA AUG UGC CUU GAU AUU, AAA GCU GAG UUG UGA GUU AUU, and ACU CAA AAC UCG CUC CUU UUU, respectively). Ambion's siNEG (catalogue number 4404020) was used as control siRNA. This negative control siRNA was selected using a modified blast to account for short sequence length and demonstrated to exclude significant homology to any known gene targets in RefSeq and MirBase (more detailed documentation on this negative control siRNA is available on the manufacturer's website [29]). Hepatocytes from female C57Black6/J mice (Charles River, Maastricht, The Netherlands) were isolated through retrograde collagenase perfusion and cultured in collagen-coated dishes exactly as previously described [30]. Twenty four hours after isolation 10 6 cells were transfected with the siRNA (final concentration 0.3, 3 or 30 nM) using the Dharmafect Duo transfection reagentH (Dharmacon, T-2010-03) according to the manufacturer's protocol. Twenty-four hours after transfection, levels of HNF4a and C/EBPa mRNAs were determined by quantitative real-time PCR (see below). Sequences from siRNAs yielding maximal reduction of transcript levels at a siRNA concentration of 3 nM were considered for use in in vivo studies.

Gene knockdown in mouse liver
Control siNEG, siHNF4a and siC/EBPa (Ambion In-Vivo-Ready for catalogue numbers 4404020, 67633 and 63855, respectively) were complexed with InvivofectamineH 2.0 Reagent (Invitrogen, Life technologies Corporation, USA) exactly according to the manufacturer's protocol. Subsequently, female C57Black/6J mice (weighing 17-19 gram) were intravenously injected via the tail vein with 200 ml complexed siRNA at a dose of approximately 7 mg of siRNA per kg body weight (in total 54 animals, 18 animals per siRNA). At two and five days post siRNA injection, animals (9 mice per siRNA for each time point) were anesthetized by a subcutaneous injection with a mixture of ketamine (100 mg/kg), xylazine (12.5 mg/kg) and atropine (125 mg/kg) after which the abdomen was opened by a midline incision and a blood sample on sodium citrate (final concentration 0.32%) was drawn from the inferior caval vein. Plasma was obtained by centrifugation and stored at 280uC until use. Liver was isolated and weighed, and liver left lobule was snap-frozen for mRNA and protein analyses and stored at 280uC until use. All mice were housed under a 12-h light/dark cycle, with standard chow diet and drinking water provided ad libitum. All experimental procedures were approved by the animal welfare committee of the Leiden University (under registration # 11005).

RNA isolation and real-time RT-PCR
Liver samples (20-30 mg) were homogenized in RNAzol (Tel-Test) and RNA isolation and cDNA synthesis was performed as previously described [31]. Gene-specific quantitative real-time PCR (QPCR) primers for Pck1, Fasn, Scd1, Lgp and Gys2 have been described previously [32]; all other gene specific QPCR primers were designed with the Primer Express software (Applied Biosystems). QPCR primer sequences are presented in Table 1. QPCR was performed on the ABI Prism 7900 HT Fast Real-Time PCR System from Applied Biosystems and data were analysed using the accompanying Sequence Detection System software. The comparative threshold cycle method with b-actin as internal control was used for quantification and normalization. siNEGinjected animals were set as a reference and the DCt values of the individual samples were related to the mean DCt of the reference group.

Plasma analyses
Plasma alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP) and bilirubin (total and conjugated) levels were determined using routine clinical chemistry assays. Global coagulability of the plasma was determined by measuring the prothrombin time (PT, Simple Simon PT system, Zafena), and the activated partial thromboplastin time (APTT) by using the STA Neoplastin Plus reagent (Roche) on the STart 4 analyzer (Diagnostica Stago). Plasma F5 activity was analyzed by using chromogenic substrate conversion [33] and activity levels of factor (F) F11 and F12 were measured with APTT-based assays [31]. Plasma fibrinogen antigen levels were assessed with a commercial murine ELISA kit from Affinity Biologicals. In plasma assays of individual coagulation factors, pooled normal mouse plasma was used to generate standard curves and the control siNEG-injected group was set as a reference (100%).

Statistical analyses
Data were analysed with the GraphPad Instat software. Statistical differences between control siNEG and siHNF4a or

siRNA screening and validation
To select an effective siRNA, mouse primary hepatocytes were transfected with three different predesigned synthetic siRNAs for HNF4a, and C/EBPa mRNAs. For HNF4a, all three siRNAs were highly and equally effective with over 95% reduction of Hnf4a transcript levels at a final concentration of 30 nM ( Figure 1A). siRNA duplex #3 had the highest level of knockdown of Hnf4a transcript levels at the 3 nM concentration and was selected for use in in vivo experiments. For C/EBPa, all three siRNAs tested were highly effective, albeit that the levels of reduction (,80%, Figure 1B) were somewhat lower as compared to the siRNAs for HNF4a. For C/EBPa siRNA duplex #1 was selected, being the most effective at 3 nM. The siHNF4a #3, siC/ EBPa #1 and the siNEG were subjected to large-scale preparation in the lipid-based in vivo transfection reagent optimized for hepatic delivery, and injected intravenously into C57Black/6J female mice. Two days after injection, siHNF4a and siC/EBPa produced a more than 75% reduction in liver Hnf4a and Cebpa transcript levels ( Figure 2A and 2B, respectively), as well as strongly reduced liver HNF4a and C/EBPa protein levels ( Figure 2C). For siC/ EBPa this strong level of knockdown of liver C/EBPa mRNA and protein persisted for at least five days ( Figure 2B and 2D). For siHNF4a, HNF4a mRNA and protein levels remained reduced at five days, but, as quantified for the transcript, only at a mean level of 36% (Figure 2A and 2D). This relatively quick return to normal HNF4a levels for siHNF4a was also observed in a second independent experiment and could not be overcome by a repeat intravenous siHNF4a injection (7 mg/kg) at day 2 after the first injection (data not shown).
During the 5-day observation period siHNF4a and siC/EBPa did not affect mouse body weight and liver weight as compared to siNEG injected animals or uninjected controls. Gross pathological analysis revealed no abnormalities. Plasma bilirubin, ALT, AST and ALP in siNEG injected animals were comparable to those observed for uninjected controls (Table 2) as expected [28]. As compared to the siNEG injected controls, at two days post siRNA injection, ALP levels were significantly increased by siHNF4a, and at 2 days significantly reduced by siC/EBPa. At 2 and 5 days, plasma bilirubin, AST, ALT levels in siHNF4a and siC/EBPa were below the limits of detection and comparable to siNEG injected animals.
Changes in transcription of control genes following liver HNF4a and C/EBPa knockdown In livers of 6-wk-old liver-specific HNF4a deficient animals [25] apolipoprotein C2 (Apoc2), apolipoprotein A4 (Apoa4), and cytochrome P450 family member 4v3 (Cyp4v3) are among the genes that are highly expressed in the liver and are strongly downregulated by prolonged HNF4a ablation, whereas in these mice scavenger receptor B1 (Scarb1) and CD36 (Cd36) are among the genes that are strongly upregulated [25,34] ( Figure 3B). The upregulation is suspected to be secondary and may result of prolonged hepatic HNF4a ablation. In livers of mice two days post siHNF4a injection, like in livers of liver-specific HNF4a deficient mice, strongly reduced Apoc2, Apoa4 and Cyp4v3 transcript levels were observed ( Figure 4A). Also at five days, Apoc2, Apoa4, and Cyp4v3 remained significantly reduced in siHNF4a compared to the siNEG injected animals, despite the limited level of HNF4a knockdown at this time point ( Figure 4C and 2A). Scarb1 and Cd36 transcript levels were not affected by siHNF4a injection at two or five days ( Figure 4A and 4C). Mice in which hepatic C/EBPa has been targeted by recombinant adenovirus encoding siRNA against C/EBPa mRNA demonstrated a role in (fasted) liver glucose and fat metabolism by affecting amongst others transcription of phosphoenolpyruvate carboxykinase (Pck1), glycogen synthase (Gys2), fatty acid synthase (Fasn), stearoyl-CoA-desaturase 1 (Scd1) and liver glycogen phophatase (Lgp) [32]. As demonstrated in Figure 4B, two days post siC/EBPa injection, the livers of (nonfasted) mice displayed significant changes in transcript levels of Pck1 and Fasn, Scd1 displayed a 40% non-significant decrease, whereas Gys2 and Lgp were unaffected as compared to siNEG injected animals. As in C/EBPa knockouts [35], siC/EBPa injection resulted in a significant increase in transcript levels of Cebpb (+73%, Figure 5) which is considered a compensatory response to C/EBPa knockdown [35]. Cebpb, Pck1, Fasn and Scd1 transcript levels in siC/EBPa-injected mice were comparable to levels in siNEG animals at five days after injection, despite the persistent strong level of knockdown ( Figure 2B and 4D).

Changes in liver coagulation gene transcription following HNF4a and C/EBPa knockdown
Livers of siHNF4a injected animals with strong reduction in HNF4a transcript and protein levels (two days) displayed markedly and significantly reduced transcript levels of Hrg (297%), Proz (270%), Serpina5 (262%), F11 (250%), F12 (246%), F13b (241%), Serpinf2 (236%), F5 (238%), F9 (227%) ( Figure 6A). For the other coagulation genes that were analysed ( Figure 3A) we did not observe significant changes as compared to siNEG injected animals. At five days, transcript levels of Hrg, Proz, Serpina5, F11 and F12 (273%, 224%, 279%, 226%, and 227%, respectively) remained reduced, despite the limited level of HNF4a knockdown at this time point (Figure 2A, 6C). Interestingly, at five days, significant elevations in transcript levels of Serpinf2, Serpind1 and Pros1 (+20%, + 22%, +32%, respectively) were observed ( Figure 6C). Overall, the HNF4a-mediated downregulation of coagulation gene transcript levels in livers of siHNF4a injected animals seemed to largely reproduce those observed in 6 week old HNF4a-null mice with prolonged HNF4a ablation in the liver from birth on (compare Figure 3A, 3B), albeit at a lower extent and, importantly, with two-day siHNF4a injected animals not showing any upregulating effects.   Compared to HNF4a, C/EBPa knockdown had a more modest impact on coagulation gene transcription. Strong reduction in C/ EBPa transcript and protein levels resulted in small though significantly reduced transcript levels of fibrinogen a and c (Fga and Fgg 225% and 224%, respectively) and F5 (227%) at two days post siC/EBPa injection. Only, the reduction in F5 transcript levels persisted upon prolonged C/EBPa knockdown (231%). At this time point also F11 transcript levels became significantly reduced (221%). For the many other coagulation genes that were analyzed, neither at two ( Figure 5) nor at five days (data not shown) post siC/EBPa injection, we observed changes as compared to siNEG injected animals that reached statistical significance.
At two days, plasma from siHNF4a and siC/EBPa injected animals had an APTT and PT comparable to siNEG injected animals ( Table 2). Although analyzed for only a limited number of individual coagulation factors, at this time point, HNF4a knockdown coincided with a significant reduction in plasma F11 activity levels, but not F12 activity levels. C/EBPa knockdown coincided with significant reduction of plasma fibrinogen antigen but not F5 activity levels. At five days, siHNF4a injected animals displayed a minimal but significant shortening of the PT, while that of siC/EBPa injected animals was minimally but significantly prolonged. Plasma fibrinogen antigen levels remained reduced upon prolonged C/EBPa knockdown.

Discussion
In the present study we used an in vivo synthetic small interfering RNA approach to determine in mice the direct role of the liver transcription factors hepatocyte nuclear factor 4a (HNF4a) and CCAAT/enhancer-binding protein a (C/EBPa) in coagulation gene transcription under conditions minimizing methodology or target-related secondary effects. Shortly (two days) post intravenous siHNF4a injection, we observed (strong) reductions in transcript levels of the Hrg, Proz, Serpina5, F11, F12, F13b, Serpinf2, F5, and F9 genes indicating that these coagulation genes are under direct regulatory control of HNF4a. The relatively modest but fast reduction in transcript levels of Fga, Fgg and F5 observed two days  [25]. Data are presented for mice two days post siRNA injection or app. 6 weeks gene ablation (i.e. 6 weeks old HNF4a-null mice). b-actin was used as internal control for quantification and normalization. The DCt values of the individual samples were related to the mean DCt of the siNEG or FLOX control group. Data are expressed as mean with error bars representing the difference between 2 POWER of upper and lower range of the mean DDCt (see figure 1 legends). Data were statistically analysed using Mann Whitney Rank Sum test. Pvalues,0.05 were regarded as statistically significant. *P,0.05, **P,0.01, ***P,0.001. On the x-axis the coagulation genes are ranked according to the magnitude of effects observed in the siRNA injected animals. doi:10.1371/journal.pone.0038104.g003 post-siC/EBPa injection indicating that these genes likely targets for direct regulatory control by C/EBPa. However, siC/EBPa injection also rapidly induced increased C/EBPb transcription, indicating a rapid onset of C/EBPa-related secondary (compensatory) effects and therefore we cannot exclude an underestimation of C/EBPa's direct role in coagulation gene transcription. Overall, we conclude that in the mouse, HNF4a has a direct and essential regulatory role for multiple hepatic coagulation genes. For C/ EBPa, such a role is more restricted, but may be underestimated as result of an unexpectedly fast compensatory upregulation of Cebpb transcription.
Analyzing the impact of liver HNF4a deficiency on a genomewide scale using microarrays identified an estimated 20% of the hepatically expressed genes affected by HNF4a deficiency to be upregulated [32,34]. Scarb1 and Cd36 were among the genes that were strongly upregulated in liver-specific HNF4a deficient mice while unaffected upon acute siRNA-mediated HNF4a knockdown in mice. It was speculated that Scarb1 increase may be a secondary consequence of altered lipid homeostasis (e.g., due to changes in intracellular lipid levels in liver of HNF4a-null mice) [25]. A similar mechanism may account for Cd36 albeit that this gene is predominantly in the liver Kupffer cells. The absence of any statistically significant upregulating effects regarding transcription in livers of mice two days post siHNF4a injection, including Scarb1 and Cd36, suggests that at least a large portion of the upregulating effects observed in the liver-specific knockouts are secondary to prolonged HNF4a disruption. This supports the use of our siRNA approach for in vivo studies on the regulating role of liver transcription factors. In this light, it is worthwhile to denote that prolonged knockdown i.e. 5 days post siHNF4a, despite a limited efficacy of hepatic HNF4a knockdown (Figure 2A and 2D), resulted in likely secondary upregulating effects for a number of the coagulation genes (Serpinf2, Serpind1, Pros1, Figure 6C).
In the present study, C/EBPa was selected as the second liver transcription factor in this first synthetic siRNA study on transcriptional control of liver coagulation genes given the observations from a recent CHIPseq study on the genome-wide occupancy of C/EBPa in livers of multiple species [16]. In this study, fibrinogen A (FGA) and prothrombin (F2) were among the few genes (from a total of 32 in mouse genome) that were located near ultraconserved C/EBPa binding regions. Indeed, in our mouse study we identified Fga as one of the few genes likely to be under direct transcriptional control of C/EBPa. However, F2 was clearly not affected following siC/EBPa, suggesting that the C/ EBPa binding site located 64 bp from the transcription start site in the mouse F2 promoter is not critical for physiological control of F2 transcription. In addition, we were surprised not to find an effect of siC/EBPa on mouse hepatic F9 transcription. In humans, carriers of hemophilia B Leiden have a causal mutation in a C/ EBPa binding site in F9 promoter [17], and in line, mice with prolonged C/EBPa ablation the liver display reduced hepatic expression of F9 [18]. Possibly, C/EBPa interaction with the F9 promoter, and also that of F2, is of high affinity, requiring only limited levels of C/EBPa binding to drive transcription. Despite the strong level C/EBPa reduction of transcript (292% and 287% at 2 and 5 days respectively) and protein levels ( Figure 2B and 2D) by our synthetic siRNA approach this may not be sufficient to unmask a role for C/EBPa for these type of targets. Alternatively, the observed unexpectedly rapid upregulation of C/ EBPb transcript following siC/EBPa injection, may functionally replace C/EBPa in liver [36]. Indeed, C/EBPb has been shown to compensate for loss of C/EBPa in the regulation of Pck1 gene expression [37]. Thus, provided the C/EBPb transcript rapidly translates to protein (which we did not determine) a compensation for loss of C/EBPa by C/EBPb cannot be excluded and may explain the absence of effects of siC/EBPa on F2 and F9 and possibly other genes.
The in vivo siRNA delivery procedure used had low toxicity without effects on circulating liver enzymes tested ( Table 2) as expected [28]). However, as compared to the control siRNA, both siHNF4a and siC/EBPa had a mild transient effect on serum alkaline phosphatase levels, the circulating marker for biliary obstruction. This suggests the presence of target-related (mild) hepatotoxic effects. Alternatively, changing in ALP levels may reflect a specific transcriptional regulatory role of HNF4a and C/ EBPa for genes involved in regulation of bile acid biosynthesis, as has been reported for HNF4a [25]. In the present study, whether siHNF4a or C/EBPa had an immediate effect on bile acid biosynthesis genes like Cyp7a1, Cyp27a1 and Cyp8b1 was not investigated.
Negative control siRNAs -siRNAs with sequences that do not target any gene products -are essential to control for the effects of siRNA delivery, and to determine whether an siRNA is considered  legends). Data were statistically analysed using Mann Whitney Rank Sum test. P-values,0.05 were regarded as statistically significant. *P,0.05, **P,0.01, ***P,0.001. Only those coagulation genes are presented for which transcript levels were significantly affected by siHNF4a or siC/EBPa (in two or five days study group). On the x-axis the coagulation genes are ranked according to the magnitude of effects observed in the two day groups. doi:10.1371/journal.pone.0038104.g006 to have a positive, negative, or neutral effect in a particular assay. In our animal studies, we included a commercially available negative control siRNA that was designed to have no significant sequence similarity to mouse, rat, or human transcript sequences (for description see methods and [29]). This negative control siRNA incorporates the same chemical modifications and is purified to the same rigorous specifications as the target-specific siRNAs (siHNF4a and siCEBPa). In addition, this negative control siRNA virtually lacked effects on gene transcription as determined for multiple cell lines following exposure of relatively high doses of negative control siRNA and analyzed by whole genome expression arrays. Despite the careful design of the negative control, we cannot exclude that in our experiments the control siRNA itself had influence on our genes of interest, and thereby leading to misinterpretation of the findings. To fully exclude such misinterpretation, in vivo experiments should be expanded with multiple carefully designed negative and multiple target specific siRNA. The HNF4a-mediated (downregulatory) changes in control and coagulation genes in livers of siHNF4a injected animals largely reproduced that of 45-day-old HNF4a-null mice and wild-type mouse primary hepatocytes rapidly after siRNA-mediated HNF4a knockdown, indicating that the single negative control siRNA approach used in the present study allowed reliable estimation of the (direct) effects of siHNF4a and also that of siC/EBPa.
In vitro siRNA screening and validation in (primary) mouse hepatocytes showed that the siRNAs targeting HNF4a had higher efficacy than those targeting C/EBPa (Figure 1). Remarkably, in vivo we observed the opposite i.e. the siC/EBPa were more effective than siHNF4a (Figure 2). Both HNF4a and C/EBPa displayed normal and stable expression in primary mouse liver cells i.e Ct comparable to fresh livers. This suggests that the discrepancy between the in vitro and in vivo findings could not be attributed to a rapid decline in HNF4a in vitro and thereby the efficacy of the siHNF4a is over estimated. Whether the difference in efficacy is due to differences in in vivo siRNA delivery, processing and or stability of the siRNA is unknown. However, it emphasizes that the in vitro experiments (as performed) are useful for identifying siRNAs with in vivo potential, but are not fully predictive for identifying most effective siRNA for in vivo use. Although, these aspects of in vivo siRNA approach should be improved, this study demonstrates that synthetic siRNA provides a simple and fast means for determining direct transcription factor involvement in vivo under conditions minimizing secondary effects. Here, in vivo siRNA-mediated knockdown enabled us to establish the direct contribution of HNF4a and C/EBPa to the regulation of coagulation gene transcription.