Targeted Gene Silencing to Induce Permanent Sterility
Author’s address (for correspondence): GA Dissen, Division of Neuroscience, Oregon National Primate Research Center, 505 N.W. 185th Avenue, Beaverton, OR 97006-3448, USA. E-mail: firstname.lastname@example.org
A non-surgical method to induce sterility would be a useful tool to control feral populations of animals. Our laboratories have experience with approaches aimed at targeting brain cells in vivo with vehicles that deliver a payload of either inhibitory RNAs or genes intended to correct cellular dysfunction. A combination/modification of these methods may provide a useful framework for the design of approaches that can be used to sterilize cats and dogs. For this approach to succeed, it has to meet several conditions: it needs to target a gene essential for fertility. It must involve a method that can selectively silence the gene of interest. It also needs to deliver the silencing agent via a minimally invasive method. Finally, the silencing effect needs to be sustained for many years, so that expansion of the targeted population can be effectively prevented. In this article, we discuss this subject and provide a succinct account of our previous experience with: (i) molecular reagents able to disrupt reproductive cyclicity when delivered to regions of the brain involved in the control of reproduction and (ii) molecular reagents able to ameliorate neuronal disease when delivered systemically using a novel approach of gene therapy.
There is an urgent need to control the exploding populations of feral cats and dogs. These animals spread disease amongst theirs and other species, are subject to abuse by humans, negatively impact human health and suffer unnecessarily pain because of malnourishment and disease. While mass sterilization via surgery is expensive and inefficient, chemical/immunological sterilization techniques mostly result in transient infertility. In addition to reducing the feral cat and dog populations, if a safe and inexpensive alternative to surgical sterilization could be provided to pet owners, this could also reduce the uncontrolled reproduction of pets that are abandoned and become the strays and unwanted cats and dogs.
A natural control point of the reproductive axis is the release of gonadotropin-releasing hormone (GnRH). In all mammals, reproductive success requires the pulsatile release of GnRH from the hypothalamus. An increase in GnRH release is a key event underlying the initiation and progression of puberty (Ojeda et al. 2010); once firmly established, GnRH pulsatility is central to the normalcy of reproductive cyclicity in females and fertility in both sexes. GnRH is released into the portal vasculature by a network of neurons that in many species – including humans, monkeys, sheep, dogs and cats – are mostly located in the medial basal hypothalamus (Barry and Dubois 1975; Belda et al. 2000; Heger et al. 2007; Dissen et al. 2012a). Neuronal (Kordon et al. 1994; Ojeda and Terasawa 2002) and glial (Ojeda and Terasawa 2002; Ojeda et al. 2003) inputs provide coordination to this network. Neurons act transsynaptically to regulate GnRH secretion using either excitatory or inhibitory neurotransmitters/neuromodulators (Terasawa and Fernandez 2001; Ojeda and Terasawa 2002; Plant and Witchel 2006). Episodic release of GnRH – a mode of hormonal output essential for reproduction – is driven by a subset of neurons located in the arcuate nucleus of the medial basal hypothalamus. These cells are called KNDy neurons (Lehman et al. 2010; Navarro et al. 2011), because they produce kisspeptin, neurokinin B (NKB) and dynorphin (Wakabayashi et al. 2010; Navarro et al. 2011). KNDy neurons release NKB, which acts on other KNDy neurons via specific receptors to stimulate kisspeptin release (Wakabayashi et al. 2010; Navarro et al. 2011). NKB and kisspeptin are released periodically, and this oscillatory behaviour is determined by a phase-delayed inhibitory feedback of dynorphin on NKB release (Wakabayashi et al. 2010; Navarro et al. 2011). Both kisspeptin and NKB have been shown to be required for normal fertility; deletion of the gene or the receptor for either peptide results in infertility (Bienvenu et al. 2000; Funes et al. 2003; Semple et al. 2005; d’Anglemont et al. 2007; Clarkson et al. 2008; Topaloglu et al. 2008; Silveira et al. 2010).
Our strategy for devising a non-surgical method to induce permanent sterility calls for the combined use of a ‘gene silencing’ technology with a ‘gene therapy’ delivery system that can target silencing molecules to specific regions of the brain critical for fertility and cause long-term suppression of gene expression. Because the vector to be employed remains active for years, permanent sterility is expected to occur following a single systemic administration.
Gene Silencing Through RNA Interference
RNA interference (RNAi) is a mechanism of post-transcriptional gene regulation in both plants and animals. In nature, the most basic unit of the RNAi system is the microRNA (miRNA) (Krol et al. 2010); scientists have produced artificial miRNAs known as small interfering RNAs (siRNAs) (Provost et al. 2002; Cullen 2005). Both the natural miRNA and the siRNA are short (∼22 base pairs), double-stranded RNAs (Cullen 2005). The miRNA is produced as part of a larger primary miRNA structure containing stem–loop regions. Processing of the primary miRNA by the RNase III enzyme Drosha releases the stem–loop pre-miRNA structure. The artificial version of the stem–loop pre-miRNA is the short hairpin RNA (shRNA). Both the pre-miRNA and the shRNA are substrates of a second RNase III enzyme known as Dicer which releases the natural miRNA or the artificial siRNA (Lee et al. 2003; Gregory et al. 2004). The cell does not distinguish between the natural miRNA and the artificial siRNA. The antisense or guide strand of the miRNA is incorporated, by base pairing to the mRNA, into a ribonucleoprotein RNA-induced silencing complex (RISC) (Khvorova et al. 2003; Schwarz et al. 2003; Cullen 2005). The next step is determined by the degree of similarity of the miRNA to the target mRNA. A high level of complementarity results in the cleavage of the mRNA resulting in suppression of transcription. Low complementarity results in imperfect binding, often to the 3′-untranslated region (UTR); this arrangement leads to translational repression and mRNA destabilization (Guo et al. 2010; Boudreau et al. 2011). It is thought that pairing with as few as 6–7 nucleotides of the miRNA is sufficient to achieve silencing (Lewis et al. 2005; Boudreau et al. 2011). Understanding how miRNAs are naturally produced and how they achieve gene silencing has allowed scientists to utilize this system to direct the silencing of virtually any gene for therapeutic purposes.
Can RNAi be Used to Disrupt Fertility?
In recently published studies of a gene know as enhanced at puberty 1 (EAP1), evidence was provided that RNAi can be used to disrupt reproductive cyclicity in two species of animals, rats and non-human primates. In an initial study carried out in female rats, Heger et al. (2007) determined the effect of a region-specific decrease in EAP1 production on the onset of female puberty and adult reproductive cyclicity. A shRNA against EAP1 mRNA was incorporated into the 3′-long terminal repeat of a lentiviral (LV) vector, which was microinjected bilaterally into the pre-optic area (POA) of juvenile 23-day-old female rats (Heger et al. 2007). The POA anteroventral periventricular nucleus (AVPV), of rats, contains most of the GnRH neurons involved in the hypothalamic control of gonadotropin secretion by the pituitary and is required for the recurrent surges of gonadotropin released during the oestrous cycle (Simerly 2002; Herbison 2006). Control animals were injected with LV particles devoid of inhibitory RNAs and containing an enhanced green fluorescent protein (eGFP)-reporter gene (Heger et al. 2007). Transduced cells were located along the lateral borders of the AVPV. The infected cells had the morphological appearance of neurons, and some were identified as GnRH neurons, by immunohistochemistry. The effectiveness of the shRNA used was shown by the reduced content of EAP1 immunoreactive material seen in infected cells vs non-infected cells.
Importantly, the time of puberty (defined by the time of first ovulation) was delayed in rats in which the expression of Eap1 gene was knocked down by shRNA in the POA as compared to LV eGFP-injected rats (Heger et al. 2007). Eap1 knock-down rats also exhibited a disrupted oestrous cycle. They had prolonged episodes in oestrus, and reduced plasma LH, FSH and estradiol levels.
The study in non-human primates was undertaken to determine whether EAP1 is also required for reproductive cyclicity in a species closer to humans. Because of improvements in RNAi technology, an artificial primary miRNA was inserted into the body of the LV vector, allowing for a more natural processing of the inhibitory RNA. The cells in the monkey hypothalamus that had been transduced by the LV constructs were also identified by immunohistofluorescence detection of eGFP (Dissen et al. 2012a). Cells positive for eGFP were found throughout the medial basal hypothalamus–arcuate nucleus region in animals receiving correctly placed injections. The transduced area showed no signs of an inflammatory reaction, as the distribution of cell populations identified by Hoechst staining of cell nuclei was normal and astrogliosis as absent (Dissen et al. 2012a). As previously observed in the rat, cells infected with LV particles carrying RNAi had a lower content of EAP1 immunoreactive material than non-infected cells.
Only normally cycling monkeys were used for this study. Control animals continued to cycle normally following the injections of LV particles containing a control miRNA. However, menstrual cyclicity was arrested in monkeys receiving the EAP1 miRNA (Dissen et al. 2012a). Animals in which the microinjections were misplaced continued to cycle after the injection.
These studies in rats and non-human primates clearly demonstrate that RNAi is capable of disrupting female reproductive cyclicity. However, the vehicle carrying RNAi had to be delivered intracerebrally to ensure effectiveness. This route is not only undesirable, but also unpractical for the induction of infertility in dogs and cats. Because one of our groups succeeded in targeting genes to the brain vasculature using a novel gene delivery method, our laboratories have joined efforts to devise an approach that can be effectively used to overcome this major limitation of gene delivery.
RNAi Delivered to the Brain via Systemic Administration of a Novel Delivery Vehicle
Although the aforementioned studies suggest that EAP1 would be a good target for RNAi-mediated silencing, EAP1 expression is widespread throughout the body (Su et al. 2004) (database accessed via http://en.wikipedia.org), thereby limiting the specificity of a systemic approach to gene silencing. An ideal target gene should be essential for fertility, exhibit tissue-specific expression and not be required for processes other than reproduction. The Kiss1 gene meets these criteria, because it is required for normal fertility in both humans and mice (Bienvenu et al. 2000; Funes et al. 2003; Semple et al. 2005; Clarkson et al. 2008), and has a limited tissue distribution, including the placenta (Su et al. 2004) (database accessed via http://en.wikipedia.org), the hypothalamus (Gottsch et al. 2004; Clarkson et al. 2009) and the ovary (Gaytan et al. 2009). Finally, mice lacking Kiss1 are infertile and exhibit no other gross abnormalities (d’Anglemont et al. 2007).
The lentiviral delivery vector we used when targeting RNAi to the hypothalamus via intracerebral injections, while effective, is not a good choice for the purpose of targeting RNAi to the brain in either domestic or wild animals. If the LV could access the hypothalamus, its genome would become incorporated into the genome of the host cell. While this is advantageous for permanent gene therapy, there is also a risk of genomic alterations that could lead to cancer or other gene mutations. A viral vector that does not incorporate into the genome would be a better choice. One such vector is the recombinant adeno-associated virus (AAV) vector. Recombinant AAV inserts its genome as a double-strained episome in the cell nucleus. Although AAVs remain episomal, they drive transgene expression for long periods of time. Adeno-associated viruses have been shown to express transgenes in different tissues throughout the lifetime of rodents (Kaplitt et al. 1994; McCown et al. 1996; Xiao et al. 1996) and have been shown to express transgenes for as long as 10 years post-injection in non-human primates (Dissen et al. 2012b). The AAV vector is derived from a non-pathogenic virus, which has contributed to its acceptance for gene therapy studies in humans (Mingozzi and High 2011).
The AAV has an additional advantage in that it can be administered via intravenous injection. The disadvantage of systemically administering the AAV is that the known serotypes do not target the brain selectively. To date, several serotypes of AAV have been characterized (as well as over 100 variants), which are defined as isolates of AAV that do not exhibit cross-reaction with neutralizing sera from any other known serotype (Wu et al. 2006). Adeno-associated virus serotypes exhibit unique tropism for different tissues throughout the body and use differing cell surface receptors to gain entry into the host cell (Wu et al. 2006). Historically, AAV2 has been the most extensively studied serotype, and hundreds of publications have detailed both pre-clinical and clinical use of AAV2 to deliver transgenes in varying tissues throughout the body (Mueller and Flotte 2008). AAV2 used for these studies infects and transduces numerous tissues throughout the body with a primary target being the liver; very little of the virus crosses the blood–brain barrier and enters the brain. The targeting of AAV2 is determined by a heparin sulphate proteoglycan (HSPG)-binding domain on the surface of the AAV capsid protein.
Targeting AAV to the Brain
To alter the tropism of AAV to specifically target the brain, the HSPG-binding domain of the capsid protein must be modified. Specifically, insertion of a peptide sequence at the arginine 588 of the capsid protein (Muller et al. 2003) has been found to alter the tropism of AAV2. The selection of the peptide sequence is critical to targeting of the virus. A method of selecting the peptide sequence that has shown great promise is known as biopanning (Work et al. 2002), wherein a random peptide library is displayed on the binding domain of a bacteriophage. This phage display library is injected into the vasculature of the target species, then the tissue of interest is collected, and the adherent phage is extracted from the tissue (Work et al. 2006; Chen et al. 2009). Chen et al. (2009) demonstrated the ability of an epitope, selected by biopanning, to direct the binding of AAV2 to the vasculature of the brain. The selected peptide epitope increased viral targeting of the brain by several orders of magnitude, and a viral reporter gene was detected in cells positive for the neuronal marker NeuN (Chen et al. 2009). This finding suggests that the virus not only targeted the brain vasculature but also entered and transduced the underlying neuropil.
Several groups have successfully altered the capsid protein by insertion of peptides at this site; generally, the mutation alters receptor binding without compromising virus viability (Muller et al. 2003; Perabo et al. 2006; Work et al. 2006). In the study by Chen et al. (2009), the entire brain was used for the biopanning experiment. In the current studies, our groups are using complementary expertise to devise a system to target a biopanning-modified AAV2 vector to the medial basal hypothalamus, the region where neurons expressing Kiss1 are located. We believe that the chances of reaching this region via the vasculature are increased because the hypothalamus is more accessible than the rest of the brain to systemically delivered molecules. The hypothalamus does not appear to be isolated by a fully functional blood–brain barrier, a feature that allows significant transfer of macromolecules from the bloodstream to the hypothalamic parenchyma (Broadwell and Brightman 1976; Herde et al. 2011).
Rational Design of siRNAs to Improve Specificity
The effectiveness of RNAi as a gene silencing tool is dependent on the ability of siRNAs to specifically suppress expression of the target mRNA. There is ample evidence of siRNAs suppressing unintended mRNAs, an effect known as off-target silencing (Chi et al. 2003; Jackson et al. 2003; Semizarov et al. 2003). Off-target silencing is thought to occur when the seed region (nucleotides 2–8 of the antisense strand) pairs with the 3′-UTR sequences of mRNAs and results in destabilization or repression of those transcripts (Birmingham et al. 2006). Such off-target silencing is even thought to result in toxic phenotypes (Fedorov et al. 2006). Boudreau et al. (2011) reported that the magnitude of siRNA off-targeting (as shown by microarray) is directly related to the frequency of seed complements (hexamers) present in the 3′-UTR transcriptome. On the basis of this and other observations, a siRNA design scheme was developed and tested which prioritizes seed specificity as a means to improve the safety profile of therapeutic RNAi sequences. This approach proved successful in identifying siRNAs that effectively silenced target gene expression, induced minimal seed-related off-targeting (known as safe-seeds) and were well tolerated in mice (Boudreau et al. 2011). Using these principles, we have designed and tested siRNAs directed against Kiss1. Ongoing studies are aimed at determining whether a siRNA specifically targeting the hypothalamic cells where Kiss1 is expressed and delivered by an AAV vector modified to target the hypothalamus will interfere with fertility.
We have shown that RNAi against a specific gene delivered to the hypothalamus via intracerebral administration disrupts female reproductive cyclicity in both rats and non-human primates. Because in this latter species ovulation was blocked, the assumption is that silencing of this single gene resulted in a state of infertility. We anticipate that devising delivery systems able to selectively, or even specifically, target the hypothalamus upon systemic administration should provide the tools to silence genes essential for reproduction in a non-invasive, effective and sustained manner in dogs and cats. We also anticipate that these studies will provide the basis for new delivery strategies to the brain for basic research purposes and emerging therapies.
Conflicts of interest
None of the authors have any conflicts of interest to declare.
This work was supported by the Michelson Foundation and by NIH grants HD-24870, HD-25123, the Eunice Kennedy Shriver NICHD/NIH through cooperative agreement HD18185 as part of the Specialized Cooperative Centers Program in Reproduction and Infertility Research and was supported by the National Center for Research Resources and the Office of Research Infrastructure Programs (ORIP) of the National Institutes of Health through Grant Number RR000163.