SEARCH

SEARCH BY CITATION

Keywords:

  • forward genetics;
  • germline mutagenesis;
  • immunity;
  • mouse

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Forward genetics and the mouse
  5. Germline mutagenesis in the mouse: practical steps
  6. Forward genetics and innate immunity
  7. Forward genetics and adaptive immunity
  8. Concluding remarks
  9. Disclosure
  10. References

The last decades have seen numerous approaches being used to decipher biological phenomena, notably the strategies we employ to defend ourselves against pathogenic attacks. From microarrays to genetics to computing technologies, all have supported a better but not yet comprehensive understanding of the pathways regulating our immune system. Limitations are notably exemplified by cases of immune deficiencies in humans that often result in high susceptibility to infections or even death, without the genetic cause being evident. To provide further insight into the mechanisms by which pathogen detection and eradication occur, several in vivo strategies can be used. The current review focuses on one of them, namely germline mutagenesis in the mouse. After describing the main technical aspects of this forward genetic approach, we will discuss particular germline mutants that have all been instrumental in deciphering innate or adaptive immune responses. Mutations in previously uncharacterized genes in the mouse, like Unc93B or Themis, have demonstrated the impartiality of forward genetics and led to the identification of new crucial immunity actors. Some mutants, like PanR1, have informed us on particular protein domains and their specific functions. Finally, certain mutations identified by this non-hypothesis-driven method have revealed previously unknown gene functions, as recently illustrated by memi, which links a particular nucleoside salvage enzyme to cell proliferation and apoptosis.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Forward genetics and the mouse
  5. Germline mutagenesis in the mouse: practical steps
  6. Forward genetics and innate immunity
  7. Forward genetics and adaptive immunity
  8. Concluding remarks
  9. Disclosure
  10. References

Upon infection, the innate arm of the immune system is our first line of defence. Its tissue-resident cells are able to directly kill pathogens. In addition, they produce cytokines and chemokines that will recruit and activate additional cells to help eradicate or at least contain the infection until the adaptive immunity is fully functional.[1] The activation of adaptive responses usually takes a few days because it is a more complex and time-consuming process. Innate antigen-presenting cells like macrophages and dendritic cells, uptake pathogens and migrate to secondary lymphoid organs where they present foreign antigens and additional signals to naive lymphocytes. This will not only activate lymphocytes but also direct their specificity.[1] For example, type I interferons produced during viral infections will induce naive CD4+ T cells to differentiate into T helper 1 (Th1) cells and will participate in the suppression of Th2 and Th17 cell-mediated responses.[2] Similarly, the response of dendritic cells to Gram-negative bacteria will lead to the production of cytokines like interleukin-6 (IL-6), IL-18 and IL-23, which will induce Th17 responses.[3] Following presentation and activation, antigen-specific lymphocytes will undergo clonal expansion and in most cases clear the infection.[4] The vast majority of these activated lymphocytes thereafter die, but some evade apoptosis to develop into functional memory cells, which react more rapidly and efficiently to a second infection with the same pathogen.[4]

The importance of our immune system is best revealed by patients with genetic aberrations and the resulting altered responses to pathogens. There is, however, a discrepancy between immune-related diseases observed in humans and our understanding of the underlying genetic deficiencies that cause them. To close this gap, a number of genetic approaches relying on the production of mutants with abnormal immune responses have been successfully used. Due to the complexity of the immune system, these genetic approaches are best used in vivo, where all molecular and cellular interactions occur under physiologically relevant conditions.

Forward genetics and the mouse

  1. Top of page
  2. Summary
  3. Introduction
  4. Forward genetics and the mouse
  5. Germline mutagenesis in the mouse: practical steps
  6. Forward genetics and innate immunity
  7. Forward genetics and adaptive immunity
  8. Concluding remarks
  9. Disclosure
  10. References

Both innate and adaptive immune responses are carefully orchestrated by the combined action of numerous genes and signalling cascades. Identifying them is therefore crucial to develop better vaccines, to find new targets to cure diseases or even to manipulate the immune system to optimize its activity. To do so in vivo, two main genetic approaches can be used. The first one, reverse genetics, relies on the disruption of genes of interest and the subsequent analysis of the mutant animal, ultimately informing on the gene function.[5] Candidate genes may be selected based on expression data obtained from DNA microarrays or because they belong to a particular family of genes of interest.[6, 7] This approach is rapid because the identity of the gene is known from the beginning. However, it is hypothesis-driven, which can present several disadvantages. (i) Even if the gene is a good theoretical candidate, it might not be involved in the specific phenomenon of interest, resulting in the absence of an observable phenotype. (ii) The complete disruption of genes required for development can be lethal, preventing the analysis of adult animals. (iii) Genetic redundancy might occur and result in the absence of phenotype. (iv) The mixed genetic background in which some mutants are generated, or the unintended deletion of non-coding regions, might affect the phenotype, independently of the actual mutation.[5, 8-10] By contrast, forward genetics is not based on any functional hypothesis because it starts with a phenotype to identify the affected gene.[11, 12] It can comprise the analysis of existing strain differences, the study of naturally occurring mutations or the screening of randomly generated germline mutations.[13-17] Because of its non-hypothesis-driven nature, forward genetics is one of the most efficient approaches to identify unknown functions or genes. This is notably illustrated by memi, a mutant in deoxycitidine kinase (dCK), which identified a crucial function for DCK in proliferation and apoptosis of peripheral lymphocytes; by the scanT mutant in Zbtb1 (zinc finger and BTB domain containing 1), which revealed a previously unknown function for this BTB-ZF (Broad complex, Tramtrack, and Bric à brac–zinc finger) member in lymphocyte development; or by Themis, which identified the role of an uncharacterized gene (E430004N04Rik) and its new protein domain during thymic development.[15, 17-19]

Ideally, complete understanding of the human immune system would require the analysis of human pedigrees where linkage analysis to map disease genes requires large families with multiple affected members. This is possible but generally restricted to inbred families where the number of available samples is often limited.[20] In addition, the analysis of immune responses is usually restricted to blood samples of already ill patients, and some phenotypes might remain unrecognized until a specific organ is analysed or a particular condition occurs (specific pathogen challenge, age, etc.).[11] To overcome these problems, forward genetics has been applied to various in vivo models including the fruit fly, the zebrafish and the mouse. All of these model animals can be genetically modified, rapidly and specifically bred, challenged with particular pathogens and kept in controlled environments, and their organs can be easily analysed. However, the mouse has been a model of choice to study human immunity because it possesses both innate and adaptive responses and is genetically closer to humans than the fish.

Germline mutagenesis in the mouse: practical steps

  1. Top of page
  2. Summary
  3. Introduction
  4. Forward genetics and the mouse
  5. Germline mutagenesis in the mouse: practical steps
  6. Forward genetics and innate immunity
  7. Forward genetics and adaptive immunity
  8. Concluding remarks
  9. Disclosure
  10. References

Mutagenesis

Traditionally, 8- to 12-week-old wild-type (wt) males (G0) of an inbred strain of choice are injected with three weekly intraperitoneal doses of 90–100 mg/kg body weight of N-ethyl-N-nitrosourea (ENU), creating up to one change per million base-pairs in their genomic DNA.[21, 22] ENU preferentially affects pre-meiotic spermatogonial stem cells, in which it creates random point mutations by transferring its ethyl group to DNA base pairs.[21, 23, 24] Due to ENU's toxicity, stem cells are partially depleted, resulting in a higher susceptibility to pathogens and a reversible sterility.[25] With the above dose of ENU, most injected males will require 10–12 weeks to recover their fertility, at which point they can be bred (Fig. 1).

image

Figure 1. Practical steps of a germline mutagenesis.

Download figure to PowerPoint

Breeding

The ENU-injected males are usually bred to two wt females to generate heterozygous G1 animals which will carry an average of 30 coding changes (Fig. 2, with only two possible breeding schemes represented).[11] These G1 can be screened for dominant mutations and if phenodeviant, used to establish a mutant pedigree. Screens for dominant mutations are therefore fast, but are usually limited to non-lethal procedures as the original phenodeviant mouse is required to establish the mutant strain. In contrast, the identification of recessive mutations takes two additional generations of breeding (Fig. 2). In a backcross scheme, G1 males are mated with wt females to produce two G2 females, which are backcrossed to their father to obtain G3 animals. Each of these G3 animals will be homozygous for one out of eight mutations present in the G1 founder.[26] In addition, it can be calculated that by screening three G3 animals per G2 daughter, 49·3% of the mutations present in the G1 founder will be screened as homozygous traits.[27] In the other breeding scheme, G3 animals are obtained by the subsequent intercross of G1 and then G2 animals.[11] A variation of this later scheme in which G0 males are bred to G1 females can be used to screen for X-linked mutations.[11]

image

Figure 2. Breeding strategies to generate mutants. Wild-type (wt) males (G0) are injected with N-ethyl-N-nitrosourea (ENU) and bred to wt females. The first-generation (G1) animals can be screened for dominant mutations of interest, or bred further to generate G3 animals homozygous for recessive mutations. To do so, G1 males can be either crossed to wt females and their G2 daughters backcrossed to the G1 founder, or G1 and then G2 animals can be intercrossed. G3 animals are subsequently screened for recessive mutations of interest. Squares: males, circles: females, no fill shape: wt, half coloured shape: heterozygous, fully coloured shape: homozygous.

Download figure to PowerPoint

Screen

In theory, any mutation can be revealed as long as a strong and reliable screen is performed. Screens can interrogate individual mice for visible phenotypes (e.g. coat colour or behaviour) or non-visible ones, relying on non-invasive analysis (e.g. blood tests) or lethal procedures (e.g. viral infections).[15, 18, 28-33] Successful screens need to be reliable (based on strong wt phenotypes to reduce the number of false negatives) and ideally follow a binomial wt/mutant distribution (to limit false positives and facilitate mapping). In addition, they usually present a rapid ‘per-mouse’ turn-over, as the higher the number of individual mice screened, the higher the number of isolated phenodeviants. For example, our laboratory has used a non-invasive blood test to isolate mutations affecting virus-specific CD8+ T-cell development and maintenance following a non-lethal viral infection.[15] We first ensured that our phenotype was very reliably expressed in our inbred strain of origin (C57BL/6J) and quantifiable in a consistent manner. We also ensured that mutants could be identified by running characterized mutants in our screen and confirming a strong binomial wt/mutant phenotypic difference. Finally, we made sure that our analysis was rapid, allowing us to screen large cohorts of mice every week. Pipelines of screens are being used by several laboratories and consortia.[34-36] In these, each individual mouse is consecutively screened for several phenotypes, usually significantly increasing the rate of mutant isolated per screened mouse.

Establishment of the mutant strain

Once the isolated mutation is confirmed to affect the germline, a stable mutant stock is established, either by directly phenotyping breeders or, if the phenotyping procedure is lethal, by systematically selecting breeding pairs with a high rate of mutant progeny.[30, 37] If this latter strategy has to be adopted, one out of 16 random G3 intercross pairs should theoretically consist of two homozygous animals, therefore providing a homozygous mutant stock.

Genotyping and phenotyping

These are usually performed in parallel, until the identity of the affected gene is required for a more thorough in-depth phenotyping. In the past, genotyping was often considered to be the daunting bottleneck of any forward genetic approach. However, with recent technological advances, genome-wide mapping of random point mutations has become much faster and easier. Traditionally, a classical positional cloning is used whereby the mutant mouse is crossed to a different inbred genetic background (mapping strain) to generate hybrid animals (H1) (Fig. 3). This mapping strain differs from the mutant strain for known markers spread over the genome. Meiotic recombination between the markers and the mutation takes place in the H1 germline and these can subsequently be backcrossed either to the mutant strain (when the phenotypic difference between mutant and wt is tight) or to the mapping strain (when the mutation is dominant). Alternatively, H1 animals can be intercrossed, providing more genetic information per analysed mouse, as each H2 originates from two recombinant germ-cells and carries less genetic material of mutant origin. H2 animals are subsequently analysed for their phenotype and genotype across the panel of chosen markers. This allows calculation of a logarithm of odds (LOD) score, which represents the probability of each individual marker to be associated with the mutation by physical linkage. A LOD score above 3 is considered significant and the markers with the highest LOD score define the region containing the culpable mutation. If this region of interest is very large or has a high gene density, it can be narrowed by analysing meiotic recombination with new markers located within the region (fine mapping). When the region of interest is small enough, the gene list is interrogated for obvious candidates and these are sequenced.[15] Alternatively, the whole genomic or coding region of interest can be captured on DNA array, enriched and sequenced.[16] This mixed mapping/sequencing approach represents a very cost-efficient and time-efficient option. Finally, whole genome sequencing has been recently used and proven successful in identifying ENU-induced mutations without having to undergo any previous positional cloning.[38] However, this approach is relatively costly and requires a significant and qualified informatics platform.

image

Figure 3. Strategies for the positional cloning of random N-ethyl-N-nitrosourea (ENU) -generated mutations. Several options are available to identify causative mutations. The newest approach comprises direct whole genome sequencing.[38] A more traditional strategy consists of outcrossing the mutant strain to another inbred strain (mapping strain). The first hybrid generation (H1) is either backcrossed to the mutant or mapping strain, or H1 animals are intercrossed, depending on the nature of the mutation. The H2 mice are subsequently used to calculate a logarithm of odds (LOD) score linking the causative mutation to a genomic region of interest. Depending on the size and complexity of this region, a second round of mapping can be performed (fine mapping), obvious candidate genes can be sequenced or the coding exons can be enriched and sequenced.[16]

Download figure to PowerPoint

Forward genetics and innate immunity

  1. Top of page
  2. Summary
  3. Introduction
  4. Forward genetics and the mouse
  5. Germline mutagenesis in the mouse: practical steps
  6. Forward genetics and innate immunity
  7. Forward genetics and adaptive immunity
  8. Concluding remarks
  9. Disclosure
  10. References

To date, the impartiality underlying forward genetics makes it one of the most efficient way to identify uncharacterized functions and unknown genes. The current efforts to generate one mutant in each gene of the genome (http://www.knockoutmouse.org/) will still require each individual mutant strain to be screened in a non-hypothesis-driven way, without limiting their analysis to a few suspected functions. One example that supports the importance of this impartiality is the discovery of the role of UNC-93B in mammalian innate immune responses to viral infections. UNC-93B was originally known to be a highly conserved endoplasmic reticulum protein regulating potassium channels and motility in Caenorhabditis elegans.[39] In the mouse, the ENU-induced ‘triple D’ mutant was isolated based on a reduced tumour necrosis factor-α (TNF-α) production upon Toll-like receptor-3, -7 or -9 stimulation.[40] The ‘triple D’ mutant strain carries a missense mutation in Unc-93b, which also affects interferon-α/β production upon viral infection.[40] Interestingly, this latter phenotype was also observed in some patients suffering from herpes simplex virus encephalitis, and who were consequently found to carry loss-of-function mutations in human Unc-93b.[41] The Unc-93b immune function could hardly have been predicted based on the protein's role in C. elegans motility, but was unravelled by a germline mutagenesis screen in the mouse. The subsequent genetic study in humans confirmed this function and eventually led to the establishment of a successful treatment regimen for these patients.

Forward genetics has also proven beneficial to understand the function of particular protein domains. Predictions based on crystallographic or electro-physical analysis are undoubtedly informative, but they will never be able to provide in vivo information.[42] One example is the dominant-negative PanR1 mutation, which carries a single nucleotide transversion in the Tnfα gene.[10] This mutation does not affect TNF-α production itself, but causes a dramatic reduction in TNF-α bioactivity by disturbing the tertiary structure of TNF-α monomers. This particular mutant therefore provides important information on a particular domain of TNF-α, region IV, and its importance during TNF-α trimer/TNF-α receptor interaction.

Forward genetics and adaptive immunity

  1. Top of page
  2. Summary
  3. Introduction
  4. Forward genetics and the mouse
  5. Germline mutagenesis in the mouse: practical steps
  6. Forward genetics and innate immunity
  7. Forward genetics and adaptive immunity
  8. Concluding remarks
  9. Disclosure
  10. References

As has been the case in the field of innate immunity, forward genetics has proven to be powerful to dissect adaptive immune responses, unravelling previously uncharacterized genes, protein domains, or new protein functions. Three examples are given below to illustrate these points.

By assessing T-cell composition in the blood, Johnson et al.[18] found a T-to-A transversion in a previously uncharacterized gene, E430004N04Rik, which was called Themis. In these mice, the early termination of the protein (Y489X) severely affects T-lymphocyte development at the transition from double-positive to single-positive cells. The study of another ENU-induced mutation in Themis (T512P) showed that the protein is not essential for myeloid and B-cell development.[19] The absence of a functional THEMIS alters the expression pattern of cell cycle, survival and metabolism regulators, eventually leading to a pro-apoptotic environment and subsequent T-cell death.[18] Interestingly, both ENU-induced mutations were found to affect a previously unidentified conserved cysteine domain named CABIT-2 (amino acids 261–521). This domain defines a new family of genes, of which one of the five members, Icb-1 might also be involved in cell-cycle regulation.[18, 43, 44] Interestingly, two additional Themis mutants obtained by reverse genetics approaches were published at the same time.[45, 46] However, because both of these mutants were generated by targeting the first exon of E430004N04Rik to completely suppress its expression, they did not identify the CABIT-2 domain or its importance for THEMIS function.

New insights into B-lymphocyte maturation and humoral immunodeficiency were recently provided by a study in which ENU-generated mutant mice were screened for the production of immunoglobulin upon immunization with a mixture of antigens.[47] Two loss-of-function mutations in DOCK8, a member of the family of Rho-Rac GTP exchanging factors, were isolated and shown to disrupt marginal zone B-cell structure and B-cell synapse function, without affecting BCR signalling.[47] Before this work, DOCK8 had only been linked to human lung cancer and mental retardation in clinical studies, and no knock-out model was available.[48, 49] These two strains therefore constitute the first DOCK8 mutant models which, in addition to unravelling the function of DOCK8 in B cells, provide good animal models to study DOCK8-deficient patients who fail to mount immune responses to vaccinations.[50, 51]

Finally, our laboratory has recently used an in vivo viral infection to screen for recessive mutations affecting CD8+ T-cell memory formation and long-term maintenance.[15] This work has notably identified a previously uncharacterized function for the protein encoded by dCK. The analysis of a complete disruption of dCK had previously demonstrated its crucial role during T and B lymphocyte development.[52] However, our work on peripheral lymphocytes has unravelled a new role for DCK during proliferation and apoptosis. We were able to show that our point mutation, which affects the very end of the DCK deoxynucleoside kinase domain, has a cell-intrinsic effect on proliferation and cell death. In addition, the chronic peripheral lymphopenia existing in mutant mice leads to a state of cell-extrinsic lymphopenia-induced proliferation. Our in vivo analysis of peripheral lymphocytes therefore unravelled a previously unknown function for dCK, linking the nucleoside salvage pathway to cell proliferation and death. In addition, by affecting the very end of the C-terminal kinase domain, our mutant will contribute to understand the role of the last 14 amino acids of DCK during lymphocyte development and peripheral life.

Concluding remarks

  1. Top of page
  2. Summary
  3. Introduction
  4. Forward genetics and the mouse
  5. Germline mutagenesis in the mouse: practical steps
  6. Forward genetics and innate immunity
  7. Forward genetics and adaptive immunity
  8. Concluding remarks
  9. Disclosure
  10. References

The random germline mutagenesis strategy combined with well-designed screens has proven paramount to the discovery of previously uncharacterized genes and to the identification of new functions of known genes or of specific protein domains. However, there are several ways in which the efficiency of the traditional germline mutagenesis can be improved. First, although variable between laboratories, the average rate at which recessive mutations have been identified is around 1 per 400–500 screened mice.[15, 30, 33, 37, 40, 47] A breeding scheme in which ENU-induced mutations are provided by both parents does increase the number of mutations per G3 animal and therefore this rate (http://mutagenetix.utsouthwestern.edu/ and ref. [11]). However, this more intensive mutagenesis scheme tends to lead to increased lethality in the G2 and G3 generations, requiring increased breeding stocks. Second, to speed up the positional cloning, ENU-injected males have been directly out-crossed to the mapping strain.[53] However, the mapping strain had to be carefully selected to avoid any interference that the mixed genetic background could have on the phenotype of interest. In addition, such mutations will irremediably be in a mixed genetic background, which in itself can have an impact on the phenotype. However, the recent validation of direct whole-genome sequencing to identify ENU-induced mutations will soon make the traditional LOD score positional cloning something from the past.[38] Third, the cost and effort required to perform a germline mutagenesis are very significant. As numerous phenotypes can be screened in a non-invasive/non-lethal manner, designing pipelines of screens in which each individual animal is successively assessed for several phenotypes is the best way to make the most of each mutant mouse. This has been successfully done at on a large scale and should be more widely performed.[34-36, 54, 55]

ENU mutagenesis is still a promising tool to find new genes and new functions. It is also undeniable that high-throughput genotype-driven approaches will create vast numbers of mutants (http://www.knockoutmouse.org), which will ultimately fill the mutant gap and help our general understanding of immunity. However, these mutants will still have to be screened in an unbiased manner if one wants to unravel new functions, and the point mutations generated by chemicals like ENU are probably still the best way to create new alleles that will inform us on specific protein domains.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Forward genetics and the mouse
  5. Germline mutagenesis in the mouse: practical steps
  6. Forward genetics and innate immunity
  7. Forward genetics and adaptive immunity
  8. Concluding remarks
  9. Disclosure
  10. References