Using Genetically Modified Rodent Models in Drug Development to Explore Target Physiology and Potential Drug Effects

The work by Berman-Booty et al ¹ in this issue of Veterinary Pathology offers a relevant case study highlighting how newer gene excision technology can be thoughtfully utilized to add to the understanding of the basic biology and potential liabilities associated with loss of function of a therapeutic target.

The use of genetically modified rodent models (GERMs), namely, gene knockouts (KO) or transgenics, is a time-honored and valuable method of deriving mechanistic insight into the role of a specific gene in biological processes and for some molecular targets has shown a strong correlation between KO phenotypes and human pharmacology, thus emphasizing the translational applicability of these models to humans. ²⁵ Conventional gene deletion technology (germline knockouts) has been the most commonly used approach. However, a significant percentage of gene mutations in conventional KOs result in embryonic lethality, and while these events may inform potential reproductive and developmental effects, they may not necessarily inform how loss of a target impacts an adult population. ^{14,15,18,20,22,24} C-terminal Src kinase (Csk) is a significant regulator of the Src family of kinases (SFK), which regulates diverse biological functions; gene deletion of Csk by conventional methods is embryonically lethal. As such, Berman-Booty et al ¹ sought to add information regarding the biological role of Csk in inflammation mediated by the SFK and characterize potential hazards associated with systemic inhibition of this target in adult mice.

The authors utilized conditional/inducible KO technology in which the gene of interest (Csk) is inactivated at a given time point postpartum by use of a tamoxifen- or tetracycline-driven Cre-loxP recombination system. ^2,17 Inducible KOs can bypass a developmental phenotype, for example, embryonic lethality or phenotypic compensation, to more closely resemble the effect of pharmacological engagement in an adult animal. ^{5,14,15,18 –20,22,24}

By developing a systemic model of Csk deletion, the authors were able to compare their findings with prior cell-based and in vivo tissue-specific models of Csk loss and add to the weight of evidence describing the role of Csk in SFK-mediated inflammation. Furthermore, the study design used by the authors demonstrates the utility of these inducible model systems to answer specific questions related to the degree (or penetrance) of knockdown and potential effects related to long-term inhibition or loss of a target. For conventional germline KO models, heterozygous animals can provide useful and relevant information related to the partial loss of gene function that could be informative in assessing the consequence of partial or full pharmacological inhibition of a target (ie, 50% or 100%). Berman-Booty et al ¹ similarly were able to evaluate incomplete loss of function by altering the length of tamoxifen treatment and thereby altering the percentage of gene loss: treatment with tamoxifen for either 3 days (with >90% gene excision) or 1 day (with ∼ 40% to 90% gene excision in various tissue). Although the histopathologic findings in the 2 different tamoxifen induction studies were qualitatively similar, the severity and extensiveness of the gastrointestinal tract and lung lesions varied depending on whether Csk excision was induced with either 3 or 1 days of IP tamoxifen administration. This type of gene dosage assessment is quite informative as an indicator of potential dose-response relationships and target organ predilection for pharmaceuticals.

While it is common practice to phenotype GERMs (ie, macroscopic/microscopic anatomic pathology, clinical pathology, and functional endpoints) to provide a single baseline snapshot of the effect of gene alteration, there is significant value in conducting longitudinal phenotypic assessments to understand long-term effects of gene alteration. This is especially relevant for the pharmaceutical industry when evaluating the potential effects of therapeutics that require long-term dosing, particularly for novel targets with little known phenotypic information. Berman-Booty et al ¹ undertook this type of time course study and determined that lung lesion progression was clearly apparent in Csk cKO mice in that the severity of the lung lesion is greater in mice euthanized on day 14 compared to mice euthanized on day 7 (after 1 IP injection of tamoxifen).

In addition to inducible gene editing technology, nuclease-based systems, such as CRISPR/Cas9 system, are a relatively recent technology that has significantly reduced the cost, time, and burden of utilizing these models. ^7,9,16,21,23 The main advantages of nuclease systems are speed and efficiency in genome-wide gene targeting. Because this technology can be readily applied in any species, it has enabled the generation of transgenic rats, which is of value since rats are the most commonly used rodent toxicology species. ¹³

As with any model system, it is important to understand the advantages and limitations of these gene editing systems to select the most suitable model and the proper use of controls. ^3,7,10 Furthermore, when characterizing KO animals, it is important to compare with other transgenic models, including in vitro cell-based systems, germline KO, tissue-specific KOs, and even human data, to help to build a weight of evidence to best understand the function of a specific gene of interest. Berman-Booty et al ¹ thoughtfully chose controls that were Cre positive mice without loxP-flanked gene (wt/wt, Cre⁺) to allow differentiation of nonspecific background lesions (ie, Cre activity) from the effects of knocking out the target gene. In addition, they compared their findings in inducible Csk KOs with previous reports in granulocyte-specific Csk KOs and in mice with mutated Hck2. These analyses allowed the authors to draw conclusions of the importance of Csk in regulating neutrophil activation, adhesion, and degranulation in vivo.

At Genentech (the authors’ institution), transgenic rodents are an integral tool used to (a) evaluate potential safety liabilities of a target, (b) inform decisions on the tractability of a target through mechanistic investigation, and (c) discern whether the adverse effects of a drug are target effects (“on-target”) or non–target effects related to the drug (“off-target”). Our strategy is to trigger the generation of these models in the early research space or hit-to-lead phase, especially for targets for which the biology is largely unknown and targets in therapeutic indications with high safety bars (ie, non–life threatening illnesses). As examples of how we interpret GERM data, when a particular toxicity is consistent with the phenotype of the KO or when a toxicity is not present in the absence of the target, these are strong signs that the toxicity is target-related. ^8,11 Alternatively, when a particular toxicity persists in the absence of the target upon dosing a compound to KO animals, this is a strong indication that the toxicity is off-target. ^6,12

GERMs are an important tool in the overall drug development strategy, with the objective being to understand the probability of therapeutic success and best characterize the potential safety liabilities of a chosen target before a candidate molecule may even exist. ⁴