Genetically engineered mouse models for hereditary cancer syndromes

Abstract

Advances in molecular diagnostics have led to improved diagnosis and molecular understanding of hereditary cancers in the clinic. Improving the management, treatment, and potential prevention of cancers in carriers of predisposing mutations requires preclinical experimental models that reflect the key pathogenic features of the specific syndrome associated with the mutations. Numerous genetically engineered mouse (GEM) models of hereditary cancer have been developed. In this review, we describe the models of Lynch syndrome and hereditary breast and ovarian cancer syndrome, the two most common hereditary cancer predisposition syndromes. We focus on Lynch syndrome models as illustrative of the potential for using mouse models to devise improved approaches to prevention of cancer in a high‐risk population. GEM models are an invaluable tool for hereditary cancer models. Here, we review GEM models for some hereditary cancers and their potential use in cancer prevention studies.

Genetically engineered mouse models are an invaluable tool for hereditary cancer model. Here we review GEM models for some hereditary cancers and their potential use in cancer prevention studies.

Abbreviations

ATM

ataxia telangiesctasia mutated

BRIP1

BRCA1 interacting protein 1

CHEK2

checkpoint kinase 2

MLH

MutL protein homolog

MRE11

meiotic recombination 11 homolog A

MSH

MutS protein homolog

NBS

Nijmegen breakage syndrome

PALB2

partner and localizer of BRCA2

PMS

postmeiotic segregation

PTEN

phosphatase and tensin

TRP53

tumor protein 53

1. INTRODUCTION

While hereditary cancers represent a minority of clinical cancer cases, the genetics and natural history of these cancer predisposition syndromes are informative regarding cancer biology in general and, in certain situations, offer molecular targets for potential cancer interventions. Indeed, some germline mutations associated with hereditary cancers can also be observed in sporadic cancers. Clinical cancer predisposition syndromes are recognized as inherited in both autosomal dominant and recessive fashions. Hereditary cancer syndromes (HCS) are characterized by early‐onset, aggressive tumor development that occurs in multiple tissue types. ¹ , ² While the number of HCS is substantial, most are rare. Studies of rare HCS, such as Fanconi anemia, ataxia telangiectasia, and xeroderma pigmentosum (XP) have contributed substantially to the understanding of DNA repair mechanisms. Many HCS are due to mutations in DNA repair pathway genes. ³ Mutations in the genes involved in homologous recombination (HR)‐ mediated DNA repair pathways underly the development of hereditary breast and ovarian cancer (HBOC) syndrome. Mutations in the nucleotide excission repair (NER) pathway results in XP. Defects in DNA mismatch repair (MMR) underlie Lynch syndrome (LS), the most common of HCS. In this review, we address HCS that have been amenable to development of genetically engineered mouse (GEM) models (GEMMs) as listed in Table 1. LS models will be a focus for discussion of potential for use in development of cancer‐preventive vaccines.

TABLE 1.

List of some hereditary cancer syndromes, the associated genes, inheritance, prevalence, and the types of cancers associated with the syndrome ^a .

Cancer syndrome	Gene	Inheritance	Prevalence	Types of cancer
Constitutional mismatch repair deficiency (CMMRD)	MLH1, MSH2, MSH6, PMS2	Autosomal recessive	>200 individuals	Lynch syndrome–associated cancer in childhood, hematologic malignancies, central nervous system tumors, and embryonic tumors
Cowden Syndrome/PTEN hamartoma tumor syndrome	PTEN	Autosomal dominant	~1:200,000	Benign and malignant tumors of the thyroid, breast, kidney, and endometrium
DICER 1 syndrome	DICER1	Autosomal dominant	1:2529 and 1:10,600	Pleuropulmonary blastoma, thyroid gland neoplasia, ovarian tumors
Familial adenomatous polyposis (FAP)	APC	Autosomal dominant	1:6850–1:31,250	Colon, gastrointestinal tract cancers
Fanconi anemia	FANCA to W	Autosomal recessive and X‐linked	~1:136,000 live births b	Acute myeloid leukemia, solid tumors—particularly of the head and neck, skin, and genitourinary tract—bone marrow failure
Hereditary breast and ovarian cancer	BRCA1, BRCA2, PALB2	Autosomal dominant	1:400–1:500	Breast, ovarian, male breast cancer, prostate, pancreatic
Hereditary diffuse gastric cancer	CDH1	Autosomal dominant	1%‐3% of gastric cancer	Diffuse gastric cancer
Li‐Fraumeni syndrome	TP53	Autosomal dominant	1:3555–1:5476	Adrenocortical carcinomas, breast cancer, central nervous system tumors, osteosarcomas, and soft‐tissue sarcomas Increased risk of other cancers also
Lynch syndrome/hereditary nonpolyposis colorectal cancer (HNPCC)	MLH1, MSH2, MSH6, PMS2, EPCAM	Autosomal dominant	1:279	Colorectal cancer (CRC) and cancers of the endometrium, ovary, stomach, small bowel, urinary tract, biliary tract, brain, skin, pancreas, and prostate
Neurofibromatosis 1	NF1	Autosomal dominant	1:2052	Neurofibroma, gliomas, malignant peripheral nerve sheath tumors
Xeroderma pigmentosum	XPA, XPC, DDB2, POLH, ERCC 1 to 5	Autosomal recessive	1:1,000,000 in the United States and Europe; 1:22,000 in Japan	Sunlight‐induced cutaneous neoplasms like basal cell carcinoma, squamous cell carcinoma, melanoma within the first decade of life

^a Source: www.genereviews.org.

^b Source: https://www.ncbi.nlm.nih.gov/books/NBK430685/.

Development of cancer is a complex, multifactorial process that involves several genetic and environmental factors. Dissecting this complex process is challenging using an in vitro system. GEMMs carrying specific germline mutations found in patients have become valuable tools to validate patient‐derived mutations underlying monogenic HCS and to inform on the biology of disease development. Besides studying genetic interactions in the development of cancers, GEMMs are also important tools for development and testing new intervention and prevention strategies that target mutations associated with hereditary cancers. Development of an agent targeting cancers with a particular mutation also extends to possible use of the agents for management of cancers with equivalent mutations in the sporadic setting. This review is intended to provide a summary of the GEMMs that have been developed for some of the most common hereditary cancers and to illustrate their use in the development of cancer interception approaches.

2. GENETICALLY ENGINEERED MOUSE MODELS

Easy genetic manipulation, smaller size, and higher rate of reproduction compared with other mammalian systems have made mouse models the most widely used animal model systems in cancer research. Among the different mouse models available for cancers, GEMMs can be considered to model the genetic, molecular, and pathophysiological nature of human tumors most accurately.

Since the development of the first transgenic mice in the early 1980s, by introducing recombinant genes (myc or v‐HRas) in the mouse germline to predispose those mice to the development of cancer, numerous GEMMs have been developed. ⁴ , ⁵ Broadly, they can be classified into two groups: gain‐of‐function models developed by transgenic, conditional transgenic, or knock‐in approaches and loss‐of‐function models generated by knockout or conditional knockout of a gene of interest.

Development of different genetically modified alleles is summarized in Figure 1. Transgenic mice are generated by inserting the DNA constructs containing the gene of interest (Figure 1B) into the genome of the zygote via nonhomologous recombination. To develop knock‐in or knockout alleles, a targeting vector containing an antibiotic resistance marker flanked by the homologous region of the gene of interest is inserted into the specific locus of the mouse embryonic stem (mES) cell genome by HR (Figure 1A). Selected mES cells are then injected into blastocysts which are then implanted into a female mouse. Site‐specific double‐strand break–inducing nuclease (eg, CRISPR/Cas9) systems generate a break in the gene of interest and induce the DNA repair system. Repair of the breaks by nonhomologous end joining introduces errors in the DNA nucleotide sequence that results in knockout of the allele. For precise insertion of any fragment of DNA or specific mutation, addition of a targeting vector containing the mutation of interest with a flanking region homologous to either side of the nuclease cut site promotes the insertion of the fragment by HR (Figure 1G).

FIGURE 1.

Different strategies to manipulate the mouse genome. A, Targeting construct containing an antibiotic marker gene (NeoR) and homology to the gene to be targeted is used to knock out a gene. Homologous recombination between targeting construct and wild‐type (WT) allele replaces one or more exons with an antibiotic marker gene that leads to generation of nonfunctional allele. B, For the activation of gene, a construct containing a promoter, ≥1 exons and introns, and the rest of the cDNA with polyA tail is inserted into the mouse genome. This generates the transgenic allele. C, Conditional transgenic allele is generated by insertion of a construct that contains a promoter, ≥1 exons and introns, and the rest of the cDNA of the gene along with polyA tail and LoxP‐STOP‐LoxP sequences between the promoter and the first exon of the gene. The STOP sequence prevents the generation of protein unless it is removed by Cre‐mediated recombination. D, Conditional knockout allele of a gene is generated by targeting a construct in which ≥1 exon are flanked by LoxP sequences. For selection of the correctly targeted allele, an antibiotic marker (NeoR) is placed in an intron. The marker is flanked by FRT sequences and FLP‐mediated recombination is used to remove the marker gene after successful targeting. This generates an allele in which ≥1 exon of a gene is flanked by LoxP sequences placed in intron sequences. E, Cre‐mediated recombination to remove the sequences that are flanked by LoxP sequences. The 34‐bp‐long LoxP sequence is shown in the right. An 8 bp asymmetric spacer sequence separates two 13 bp palindromic sequences. Arrows indicate Cre‐mediated cleavage sites during recombination. Recombination between two LoxP sequences leads to the loss of the regions that are placed in between those sequences. F, In the doxycycline inducible gene activation system, the promoter driving the gene of interest is fused with tetO (tet‐operator) sequence, and a transgene driving the expression of transactivator rtTA (reverse tetracycline controlled transactivator) is inserted into the genome. Only in the presence of doxycycline, rtTA binds to tetO, leading to the transcription of the gene. G, CRISPR‐Cas9 mediated genetic engineering. A large deletion in a gene can be achieved by using two guide RNAs (sgRNA) to induce double‐strand breaks (DSB) at preferred locations of a gene. Repair of induced DSBs by nonhomologous end joining results in generation of a knockout allele. In the presence of the homologous DNA repair template, the repair of induced DSBs generates knock‐in allele. Colored boxes denote exons, and thick lines represent introns of a gene. Numbers in the boxes represent exon numbers, the promoter is marked by “P,” and arrows indicate gene transcription. Dotted crossed lines are used to represent homologous recombination.

Many transgenic or knockout mice developed by activation/deletion of genes in all cells of the body have provided important insights into the role of the gene in the tumorigenesis process. However, embryonic lethality occurs when certain genes are deleted or constitutively overexpressed in germline. This results in difficulty in the study of the role of the genes in tumor development in adult mice. To circumvent this, powerful “Cre‐Lox” technology has been developed that allows splicing out of target genes that are flanked by LoxP sequences by the Cre recombinase (Figure 1D,E). One of the first examples of conditional GEMMs was the deletion of the Apc gene in intestinal cells by adenovirus‐mediated delivery of Cre resulting in the development of colorectal adenomas that share many features of adenomas developed in familial adenomatous polyposis (FAP) patients. ⁶ In conditional transgenic mice, the inserted gene is inhibited by insertion of a LoxP‐flanked “STOP cassette” in the upstream of the target gene (Figure 1C). Cre‐mediated recombination removes this cassette, permitting the expression of the gene. This technology can also be used to insert a fragment of DNA into the genome.

Another type of GEMM in which gene activation/deletion is affected in a controlled way is the inducible GEMM. In those models, expression of Cre recombinase is controlled by small molecules or drugs. In the tamoxifen‐inducible Cre system (Cre‐ER), Cre is fused with a mutated estrogen receptor (ER) that translocates into the nucleus to induce recombination only when it binds to tamoxifen. Another system is the tetracycline‐ or doxycycline‐inducible system, in which the addition of these drugs induces the expression of Cre (Figure 1F).

Conditional deletion/activation of genes in the new‐generation GEMMs has revolutionized the use of GEMMs in prevention research, particularly for the identification of biomarkers in premalignant and malignant lesions for early detection and intervention, investigating the progress of premalignant lesions to malignant tumors, and preclinical assessment of new agents in cancer prevention. However, the success of the use of GEMMs in prevention research depends on the appropriate choice of model, such that it has biological and pathological relevance to human cancer, and on the experimental design of the study.

To study gene interactions as seen in human tumors, compound animal models have been generated by intercrossing different mutant strains to combine different mutations within the same cells. Examples of different breeding strategies are outlined in Figure 2A‐C. The tumors in the GEMMs are expected to develop similar anatomical distribution and functional consequence to human tumors. GEMMs of tumorigenesis have been used in validating candidate cancer genes, studying the tumor microenvironment and cancer‐originating cells, validating drug targets, and evaluation of drug resistance, as summarized in Figure 2D. ⁵ Also, the tumors generated in immunocompetent GEMM model the response of immune cells and are useful when evaluating immunotherapies.

FIGURE 2.

Generation of different genetically engineered mouse models and their applications. A, Breeding of mice harboring knockout allele in tumor suppressor gene to generate homozygous, heterozygous, and wild‐type cohorts for tumor studies. B, Crossing mice homozygous for conditional knockout of any gene (gene X) with Cre‐expressing transgenic mice leads to generation of compound heterozygotes that are interbred to generate experimental cohorts. Transgenic mice expressing Cre from a tissue‐specific promoter develop tissue‐specific knockout of a gene. C, Breeding of mice to generate inducible conditional knockouts. Transgenic mice with rtTA (reverse tetracycline‐controlled transactivator) transgene and transgenic mice with tetO‐regulated promoter–driven Cre expression are crossed to generate double‐transgenic mice expressing rtTA and Cre. This double‐transgenic mouse is further crossed with homozygous conditional knockouts for any gene (geneX) to generate compound heterozygotes that are interbred to generate cohorts. Addition of doxycycline induces Cre expression that further promotes deletion of the conditional allele. Solid boxes and lines represent exons and introns, respectively. “X” is used to mark deletion and “P” is used to mark promoter. Solid triangles represent loxP site. The probability of obtaining a genotype from a cross is indicated in front of the genotype. D, Schematic summary of applications of genetically engineered mouse (GEM) models to understand basic biology (top half of figure) or in translational oncology (bottom half of figure). The scheme was created using Biorender (http://Biorender.com).

3. DIFFERENT GEMMS FOR SOME HEREDITARY CANCERS

There are more than 100 cancer syndromes that are classified as hereditary (www.genereviews.org). Many of these hereditary cancers are extremely rare but all are associated with an aberration in a specific gene. ⁷ GEMMs have been generated that carry precise mutations as found in patients, for many of these genetic disorders. We describe here GEMMs for two of the most prevalent hereditary cancers: hereditary nonpolyposis colorectal cancer (HNPCC)/LS and HBOC (Table 1). GEMMs for FAP, Li‐Fraumeni syndrome, hereditary diffuse gastric cancer, DICER1 syndrome, PTEN hamartoma syndrome, neurofibromatosis 1, Fanconi anemia, and xeroderma pigmentosum are discussed in Appendix S1.

4. HEREDITARY NONPOLYPOSIS COLON CANCER/LS

Hereditary nonpolyposis colon cancer (now most commonly known as LS), characterized by early age at diagnosis of colon cancer and a defined spectrum of extracolonic cancers and transmission as an autosomal dominant trait, is the most common cancer syndrome in humans. LS patients demonstrate a heritable defect in the DNA MMR pathway and 70%‐80% of families show germline mutations in one of four MMR genes, MLH1, MSH2, MSH6, and PMS2. ⁸ , ⁹ , ¹⁰ , ¹¹ In the MMR pathway, mismatch is first recognized by MSH2‐MSH6 or MSH2‐MSH3 heterodimers depending on the nature of mismatches. Next, the MLH1‐PMS2 heterodimer generates nick on the hemimethylated DNA by PCNA/replication factor C (RFC)‐dependent endonuclease activity, which plays a critical role in the exonuclease‐mediated removal of mismatched bases and DNA resynthesis by Polymerase δ (Figure 3A). ¹² To date, majority of the patients with LS have mutations in the MLH1 or MSH2 gene, followed by a small percentage with mutations in MSH6 and PMS2. There is not yet any definite association of other MMR genes such as MSH3, MSH4, MSH5, MLH3, and PMS1 with LS. ¹³ , ¹⁴ Mutations in both the alleles of MMR genes causes a rare disorder, the constitutional mismatch repair deficiency (CMMRD) syndrome that greatly increases the risk of developing cancer in childhood (before age 18). The tumor spectrum for CMMRD mainly includes hematologic, brain/central nervous system, gastrointestinal (GI), and other malignancies. ¹⁵ Mutations in PMS2 is the most common cause of CMMRD, followed by MLH1, MSH2, or MSH6 genes. The tumor spectrum and age of malignancy onset may differ in patients with biallelic germline mutations of different MMR genes. ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹

FIGURE 3.

DNA repair pathways involved in Lynch syndrome (LS) and hereditary breast and ovarian cancer (HBOC). A, Mismatch repair pathway (MMR) in which mismatched bases are recognized by the MSH2‐MSH6 heterodimer and daughter strand is cleaved by MLH1‐PMS2–mediated endonuclease activity using assistance from the PCNA‐RFC complex. A segment of newly synthesized DNA is removed by exonuclease, and polymerase δ resynthesizes the DNA resulting in the repair of the mismatched DNA. B, Homologous recombination–mediated DNA repair. MRN complex (MRE11‐RAD50‐NBS1) recognizes the double‐stranded breaks (DSBs). BRCA1‐CtIP–mediated nuclease activity resects the DNA end from 5′ to 3′ and leads to the formation of single‐strand DNA (ssDNA) that gets coated with DNA replication protein A (RPA). Then, RAD51 nucleoprotein filament assembled by the multiprotein complex that involves BRCA2, PALB2, BRCA1 along with many other proteins replaces the RPA‐coated ssDNA, performs homology search, and mediates strand invasion. New synthesis of DNA using sister chromatid DNA as template followed by ligation and resolution of Holliday junction repairs the DSBs. Biorender (http://Biorender.com) was used to create the illustration.

The first gene that was implicated in LS is MSH2, ²⁰ and it was also the first LS gene to be inactivated in mice. ²¹ , ²² Different mouse models have been developed since then and are summarized in Table 3. Msh2‐deficient mice are viable and fertile but have a reduced life span due to the development of lymphomas, specifically T‐cell lymphomas. ²² A knock‐in mouse model (Msh2 ^G674A/G674A ) carrying a missense mutation in the conserved ATPase domain showed a similar tumor spectrum to Msh2‐null mice, though they survived longer than Msh2‐null mice. ²³

Msh2 deficiency in immunocompromised Tap1 mutant mice survive longer than only Msh2 deficient mice and develop GI and skin tumors. ²⁴ Msh2 heterozygous mice in combination with the Min allele of the Apc tumor suppressor gene (Apc ^+/Min ; Msh2 ^+/− ) develop intestinal tumors, and a significant portion of tumors showed loss of the wild‐type Msh2 allele. ²⁵ A conditional mouse model, Msh2 ^loxp/loxp ; Vill‐Cre mice, in which Msh2 is specifically deleted in the intestinal epithelium using Vill‐Cre, develops intestinal tumors and avoids lymphoma development. ²⁶ Tumors develop in Msh2 ^loxp/loxp ; Vill‐Cre mice over the course of more than 1 year in the small intestine, and tumors showed high microsatellite instability (MSI), a marker for MMR deficiency. ²⁶ Microsatellites (MS) are short tandem repeats of two to six nucleotides distributed over the genome. MMR deficiency fails to repair errors at these sites introduced during DNA replication, resulting in change of the length of MS sequences and is referred to as MS instability. According to the National Cancer Institute guidelines, MSI in tumor tissue is measured by comparing the repeat length at five different loci (BAT‐25, BAT‐26, D2S123, D5S346, and D17S250) with normal DNA. Instability at two or more loci is considered MSI high. ²⁷ To detect MSI high, next‐generation sequencing coupled with computational algorithms has been developed recently to detect multiple MS alterations in several mutational hotspots of oncogenes or DNA repair genes. ²⁸

The Msh2 ^loxp/loxp ; Vill‐Cre model for LS has been used in many preclinical studies including chemopreventive agent testing. Treatment with aspirin or naproxen increased the lifespan of Msh2 ^loxp/loxp ; Vill‐Cre mice compared with control (Figure 3A,B). ²⁹ , ³⁰

The alteration of MS sequences in the coding region of cells deficient in MMR generates frameshift peptide (FSP)‐based neoantigens. Immunogenic MSI‐associated FSPs promote cytotoxic T‐cell infiltration into the tumors. ³¹ A phase I/IIa clinical trial of a vaccine against recurrent FSP‐based neoantigens that are shared in MMR‐deficient tumors showed immunologic efficacy and safety of using a neoantigen‐based vaccine as immunologic intervention for MMR‐deficient colorectal cancers. ³² A preclinical cancer‐preventive study using Msh2 ^loxp/loxp ; Vill‐Cre mice showed that vaccination using four shared FSP neoantigens reduces tumor burden and prolongs survival (Figure 4A). ³³ This study also showed that the combination of chemo‐ and immuno‐prevention using naproxen or aspirin and FSP vaccination is synergistic and increased the survival of Msh2 ^loxp/loxp ; Vill‐Cre mice by reducing tumor burden (Figure 4A). ³³

FIGURE 4.

Preclinical use of genetically engineered mouse (GEM) models for Lynch syndrome (LS). A, Vaccination of Msh2 ^loxp/loxp ; Vill‐Cre mice with recurrent frameshift peptide (FSP) antigens prolong survival. Survival curve of Msh2 ^loxp/loxp ; Vill‐Cre mice after treatment with aspirin (ASA), naproxen (NAP), FSP antigens, ASA + FSP antigen, or NAP+FSP antigens. Untreated mice were used as control. Six‐ to eight‐week‐old mice were enrolled in the study. A total of 50 μg of each FSP combined with 20 μg of the adjuvant were introduced subcutaneously to either the right or left flank four times biweekly, followed by four times monthly. At the age of weaning, 400 ppm ASA or 166 ppm NAP was incorporated into the diet of the mice assigned to those arms of the study. ³³ B, Survival curve of Mlh1 ^−/− mice with and without vaccination using cell line lysates. Two Mlh1 ^−/− cell lines (328 and A7450 T1 M1) lysates were injected subcutaneously. Application was initiated in 8‐10‐week‐old mice at the dose of 10 mg/kg body weight by 4‐weekly administration and then monthly administration for up to 12 vaccinations (n = 9/group). Control mice (n = 15) were left untreated. *** < 0.001 A7450 T1 M1 vs. control; Log‐rank (Mantel‐Cox) test. ⁴⁰

When the Msh2 conditional allele (Msh2 ^loxp ) was combined with a knock‐in allele Msh2 ^G674D , the resulting mice, upon expression of Cre‐recombinase from intestinal epithelium (Vill‐Cre), showed increased intestinal tumor formation and high MSI. ²⁶ In another model, a conditional Msh2 allele (Msh2 ^flox ) was generated, in which exons 12 and 13 are inverted upon Cre‐LoxP‐mediated recombination. Mice (Msh2 ^−/flox ) with a combination of the Msh2 knockout allele (Msh2 ⁻ ) and Msh2 ^flox allele generated intestinal adenocarcinoma when Cre‐recombinase was expressed from the Lgr5 promoter in self‐renewing intestinal stem cells to induce Msh2 inactivation. ³⁴ , ³⁵ These results show that in combination with a lymphoma‐suppressing condition, Msh2‐deficient mice can serve as a model for LS. However, the GI tumors developed in Msh2‐deficient mice are mainly in the small intestine as compared with the colon in human LS patients.

Targeted inactivation of Mlh1 in mice reduces life span and shows a high degree of cancer predisposition that includes tumors in the GI tract, lymphomas, and a number of other organ‐site tumors. ³⁶ , ³⁷ , ³⁸ , ³⁹ Combining Mlh1 mutation with Apc1638N mutation dramatically reduced the viability of mice with a significant increase in tumor numbers in the GI tract. The GI tumors that were found in Mlh1 ^−/− mice were adenomas to early invasive carcinomas and were present throughout the small intestine. ³⁸ Prophylactic vaccination using cell lysates from Mlh1 ^−/− cell lines in the Mlh1 ^−/− mice model prolonged the tumor‐free survival and showed that cell line–derived neoantigens can induce immune stimulation in the preclinical mouse model (Figure 4B). ⁴⁰ Lysates of two cell lines (328 and A7450 T1 M1) generated from Mlh1 ^−/− spontaneous GI tumors were used for vaccination. The two cell lines had different nonsynonymous single‐nucleotide variants (SNVs) in the preselected mutational hotspots. Only two genes Arid1a and Apc are shared hotspots that showed alterations from both cell lines. Vaccination with A7450 T1 M1 lysate showed a better response compared with 328 lysates though the 328 cell lines had higher mutational burden than A7450 T1 M1. The secreted cytokine profile varies between the two cell lines, and A7450 T1 M1 cell line secretes favorable cytokines for colorectal cancers (GM‐CSF, IL‐1b). This suggests that the ability to induce favorable T‐cell response is more crucial than the number of mutations. ⁴⁰

Besides those knockout Mlh1 mice models, a conditional Mlh1 (Mlh1 ^flox/flox ) mouse model was generated in which exon 4 can be deleted using Cre‐recombinase. These conditional mice (Mlh1 ^flox/flox ) when crossed with mice expressing EIIa‐Cre allowed constitutive inactivation of Mlh1, and the resulting mice displayed a tumor predisposition like Mlh1 ^−/− mice. ⁴¹ A knock‐in mouse model of Mlh1 (Mlh1 ^G67R/G67R ) results in DNA repair deficiency due to mutation in one of the ATP‐binding domains of Mlh1 and displays a strong cancer predisposition phenotype with significantly fewer intestinal tumors. ⁴²

The MutS homolog (MSH)2 forms a complex with MSH3 or MSH6 to recognize different DNA mismatches. Genetic studies suggest that MSH3 and MSH6 partially substitute for each other in MMR of DNA. ⁴³ The MSH2‐MSH6 heterodimer recognizes a single‐base mismatch or single‐base insertion/deletion, while the MSH2‐MSH3 heterodimer primarily recognizes an insertion/deletion loop (2‐12 nucleotides). ⁴³ , ⁴⁴ , ⁴⁵ Msh6 ^−/− mice have reduced life spans but live longer than Msh2 ^−/− mice and develop a spectrum of tumors including GI tumors and lymphomas. ⁴⁶ Msh6 ^−/− mice also develop endometrial cancer. ⁴⁴ The tumors developed in Msh6 ^−/− mice do not show MSI, like human patients with MSH6 mutations that showed atypical LS with late onset of cancer development and variable MSI phenotypes. ⁴⁷ Redundant roles of MSH3 and MSH6 in MMR could explain the less extensive MSI in Msh6 ^−/− mice. ⁴⁸ Combining Msh3 ^−/− mutations along with Msh6 ^−/− accelerates intestinal tumorigenesis with tumor predisposition indistinguishable from Msh2 ^−/− or Mlh1 ^−/− mice. ⁴⁴ , ⁴⁹ This suggests cooperation of Msh6 and Msh3 in tumor suppression. ⁴⁴ Msh6 with p.Thr1217Asp mutation has a dominant effect on MMR, and a knock‐in homozygous mouse model with this missense mutation (Msh6 ^{T1217D/T1217D} ) showed similar life span and tumor spectrum but increased MSI compared with Msh6 ^−/− mice, modeling the variable MSI phenotypes found in patients with MSH6 mutations. ⁵⁰ The p.Thr1217Asp mutation resists the ATP‐induced release of the Msh2‐Msh6 T1217D complex from mismatched bases. The longer half‐life of the Msh2‐Msh6 T1217D complex interferes with Msh2‐Msh3 function in repairing dinucleotide insertion/deletion and thus increases the instability at dinucleotide repeats in Msh6 ^{T1217D/T1217D} mice, unlike Msh6 ^−/− , where the lack of Msh6 protein does not interfere with Msh2‐Msh3 function. The Msh2‐Msh6 complex is also involved in DNA damage response, and cells deficient in Msh6 or Msh2 are resistant to cisplatin or MNNG treatment. Cells homozygous for Msh2 ^T1217D mutation showed sensitivity to cisplatin or MNNG suggesting that the Msh6 T1217D does not affect DNA damage response. Similar life span and tumor spectrum of Msh6 ^{T1217D/T1217D} and Msh6 ^−/− mice suggests that Msh2 functions both in DNA repair and DNA damage response cooperate in the tumorigenesis process. ⁵⁰

Homozygous Pms2 mutant (Pms2 ^−/− ) mice are viable but infertile and showed an increased level of MSI. These mice developed lymphomas and sarcomas but did not show increased intestinal tumorigenesis. ³⁹ , ⁵¹ , ⁵² , ⁵³ However, Pms2 ^−/− mice in Apc ^Min background showed a significant increase in intestinal adenoma formation. ⁵⁴ An endonuclease‐deficient (Pms2 ^E702K/E702K ) knock‐in mouse showed increased incidences of lymphomas but failed to show an increased level of intestinal tumor formation. ⁵⁵ We have recently reported a Pms2 knock‐in mouse model (Pms2 ^{c.1993A>G/c.1993A>G} ) that harbors an equivalent mutation (NM_000535.5:c.2002A>G) as found in a Canadian Inuit population that is associated with development of late‐onset LS. This mouse model shows a defect in splicing of exon 11 and thus generates a frameshift mutation and truncated protein (I665*). Homozygous mutant mice showed elevated levels of MSI and intestinal tumor burden in Apc ^tm1Rak (a chain‐terminating mutation at codon 1638 in exon 15 of Apc) mutant background. ⁵⁶ However, both the Pms2 knock‐in mouse models are fertile unlike Pms2 ^−/− mice. ⁵⁵ , ⁵⁶

The Pms2 ^{c.1993A>G/c.1993A>G} is a valuable model to evaluate interventions designed to correct the splicing defect as a strategy for preventing development of intestinal tumors. Numerous small molecule modulators of splicing have been identified. ⁵⁷ This preclinical mouse model is well‐suited for testing the efficacy of small‐molecule splice modulators in suppressing the splicing defect and thus intestinal tumor formation.

5. HEREDITARY BREAST AND OVARIAN CANCER

Hereditary breast and ovarian cancer syndrome is an autosomal dominant cancer syndrome that is characterized by a high risk of developing breast (before the age 50 years) and ovarian cancer and an increased risk of developing other cancers, such as male breast cancer, prostate cancer, pancreatic cancer, and melanoma. Two main genes, breast cancer susceptibility gene 1 (BRCA1) and 2 (BRCA2) are linked with HBOC but other genes like TRP53, PTEN, ATM, CHEK2, BRIP1, and PALB2 are also linked with increased risk of developing breast and ovarian cancers. ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² Genes involved in HBOC play a major role in the HR‐mediated DNA repair pathway. The DNA double‐strand break is repaired without any error by the HR‐mediated repair pathway. The double‐strand DNA break is recognized by the MRN complex, which consists of MRE11, RAD50, and NBS1. The MRN complex activates ATM and ATR, kinases that phosphorylate downstream targets and initiate DNA damage response. CtIP, ubiquitinated by BRCA1, next resects the DNA to form 3′ overhangs that are bound by RPA. BRCA2 with the assistance of PALB2 and BRCA1, loads RAD51 onto RPA‐coated DNA. The RAD51 nucleoprotein filament then invades the homologous DNA strand of the sister chromatid and allows the resynthesis of DNA with the use of the sister chromatid as a template, allowing for error‐free repair (Figure 3B). ⁶³

Homozygous deletion of Brca1, Brca2, or Palb2 using different gene‐targeting strategies leads to embryonic lethality, and germline coinactivation of Trp53 delays embryonic lethality for a few days. ⁶⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷ To overcome embryonic lethality and to study the effect of loss of these genes on tumor formation, conditional mutant mouse models or models with partial loss of gene function have been developed. Homozygous deletion of Brca1 exon 11 leads to embryonic lethality at E12.5‐E18.5 dpc, and this lethality could be overcome in the presence of a p53 mutant background. The Brca1 ^Δ11/Δ11 ; p53 ^+/− mice are viable and develop mammary tumors at the age of 6‐10 months. ⁶⁸ Deletion of exon 11 in the mammary glands using Wap‐Cre or MMTV‐LTR–driven Cre results in mammary tumors with long latency (10‐13 months). The latency of tumor formation could be reduced in p53 ^+/− background. ⁶⁹ In 2001, Ludwig et al developed a mouse model (Brca1 ^tr/tr ) that generates a truncated Brca1 protein (924 aa). Homozygous mice (Brca1 ^tr/tr ) are viable and developed a variety of tumors including late‐onset breast carcinomas. This study showed that Brca1 mutations can lead to mammary tumor development without additional genetic manipulations. ⁷⁰

Mice carrying truncating mutations of the exon 11 of Brca2 are born at sub‐Mendelian ratio and show varied frequency of survival depending on genetic background. Homozygous mice that survive show a high incidence of thymic lymphomas. Similar analyses have been carried out in mouse Brca2 models suggesting that truncating mutations of the exon 11 of Brca2 is a protumorigenic event. ⁷¹ , ⁷² Many other mouse models with partial loss of function have been generated for both Brca1 and Brca2 and are listed in Table 2.

TABLE 2.

Genetically engineered mouse models for Lynch syndrome and hereditary breast and ovarian cancer (HBOC).

Disease	Gene	Allele	Type of model	Exons or domain affected	Cre and target tissue	MSI	Phenotype	Ref.
Lynch syndrome	Msh2	Msh2 loxp	Conditional	Exon 12	Villin‐Cre; intestine	Yes	Small intestinal tumor, reduced lifespan (12 mo)	26
	Msh2	Msh2 loxp	Conditional	Exon 12	Ella‐Cre; all tissues	Yes	Lymphoma, GI adenosarcoma, reduced lifespan (6 mo)	26
	Msh2	Msh2 loxp	Conditional	Inversion of exons 12 and 13	Lgr5‐CREERT2; Lgr5‐expressing stem cells		Small intestinal tumor	34
	Msh2	Msh2 G674A	Knock‐in	ATPase domain	NA	Yes	Lymphoma, GI adenosarcoma, reduced lifespan (12 mo)	23
	Msh2	Msh2 G674D	Knock‐in	ATPase domain	NA	Yes	Lymphoma, GI adenosarcoma, reduced lifespan (12 mo)	26
	Msh2	Msh2 −	Knockout	Corresponding human exon 11	NA	Yes	Lymphoma, reduced lifespan (6 mo)	21
	Msh2	Msh2 −	Knockout	Disrupts before ATP‐binding domain	NA	Yes	Lymphoma, sarcoma, reduced lifespan (6 mo)	22
	Mlh1	Mlh1 flox	Conditional	Exon 4	Ella‐Cre; all tissues	Yes	Lymphoma, intestinal and skin tumor, reduced lifespan (12 mo)	41
	Mlh1	Mlh1 G67R	Knock‐in	ATP‐binding domain	NA	Yes	Lymphoma, intestinal and skin tumor, reduced lifespan (12 mo), infertile	42
	Mlh1	Mlh1 −	Knockout	Exon 4	NA	Yes	Lymphoma, intestinal and skin tumor, sarcoma, reduced lifespan (12 mo), infertile	36, 39
	Msh6	Msh6 T1217D	Knock‐in	Msh2‐Msh6 heterodimer interface	NA	No	Lymphoma, endometrial cancer, reduced lifespan (12 mo)	50
	Msh6	Msh6 −	Knockout	3′ of exon 4	NA	No	Lymphoma, endometrial cancer, reduced lifespan (12 mo)	44
	Pms2	Pms2 E702K	Knock‐in	Endonuclease domain	NA	Yes	Lymphoma	55
	Pms2	Pms2 c.1993A>G	Knock‐in	Splicing of exon 11, generation of new splice acceptor site	NA	Yes	Increased intestinal tumor when in combination with Apc tm1Rak/+	56
	Pms2	Pms2 −	Knockout	Exon 2	NA	Yes	Lymphoma, sarcoma, reduced lifespan, males are infertile	39, 51
HBOC	Brca1	Brca1 Δ11	Conditional	Exon 11	EIIa‐Cre; All tissues	NA	Survive in p53 +/− or p53 −/− background, develop mammary tumor in 6‐10 mo	68
	Brca1	Brca1 Δ11	Conditional	Exon 11	Wap‐Cre; mammary gland	NA	Develop mammary tumor with long latency (10‐13 mo)	69
	Brca1	Brca1 Δ11	Conditional	Exon 11	MMTV‐LTR‐Cre; mammary gland	NA	Develop mammary tumor with long latency (10‐13 mo) and latency decrease in p53 mutant background	69
	Brca1	Brca1 flex2	Conditional	Exons 1 and 2	Wap‐Cre; mammary gland	NA	Increased mammary tumor formation with long latency (512 d)	77
	Brca1	Brca1 Δ5‐13	Conditional	Exons 5 to 13	K14‐Cre; Epithelial cells	NA	Mammary and skin tumor development, reduced latency (213 d) in p53 fl/fl background	78
	Brca1	Brca1 Δ22‐24	Conditional	Exons 22 to 24	Blg‐Cre; mammary gland	NA	Develop mammary gland tumor with long latency (12‐14 mo), in p53 +/− background more incidences of tumor with shorter latency (6‐46 wk)	79
	Brca1	Brca1 tr	Truncation due to frameshift	Truncation at exon 11 after 924 aa due to insertion of 50 bp	NA	NA	Lifespan reduced (1.4 y) and development of lymphoma, sarcoma, breast carcinoma/adenoma	70
	Brca1	Brca1 FL	Specific transcript deletion	Specific deletion of Brca1 Δ11 transcript and leaves full‐length Brca1 intact by inserting exon 11‐12 cDNA at exon 11 and disrupting intron 11	NA	NA	Gynecologic hyperplasia, mammary hyperplasia in adult mice, thymic lymphoma	80
	Brca1	Brca1 FH‐I26A	Knock‐in	Ring domain, mutate E3 ligase activity	NA	NA	No increased tumor development mice are normal	81
	Brca1	Brca1 S1598F	Knock‐in	BRCT domain, disrupts phospho‐protein recognition domain	NA	NA	Increased tumor incidence when combined with p53 mutation, males are sterile	81
	Brca1	Brca1 S971A	Knock‐in	Chk2 phosphorylation site	NA	NA	Uterus hyperplasia and ovarian abnormalities, increased tumor formation (lymphoma, mammary, and colon) on DNA‐damaging agent treatment	82
	Brca1	Brca1 S1152A	Knock‐in	ATM‐dependent phosphorylation site	NA	NA	Increased tumor formation (lymphoma, breast, and liver) upon radiation. Low percentages (36%) show aging‐like phenotype like growth retardation, skin abnormalities, delay in mammary gland morphogenesis, mammary gland abnormalities (18 mo old)	83
Brca1	Brca1 LP	Knock‐in	L1363P, disrupts PALB2 interaction	NA	NA	Born in Mendelian ratio but develop Fanconi anemia (FA)‐like phenotype including growth retardation, hyperpigmentation, cerebellar hypoplasia, infertility, bone marrow failure, and T‐cell lymphoblastic lymphoma	84
	Brca1	Brca1 cc	Knockout (CRISPR‐Cas9 based)	Coiled‐coil domain deletion, disruption of interaction with PALB2	NA	NA	Born in sub‐Mendelian ratio, live mice are infertile and develop FA‐like phenotype, T‐cell acute lymphoblastic leukemia, phenotypes are rescued in compound heterozygosity with Brca1 Δ11 allele	85
	Brca1	Brca1 loxp/loxp ;Trp53 loxp/loxp ;Pten loxp/loxp	Compound conditional	Exon 11 of Brca1	Pax8‐rtTA; TetO‐Cre; Doxycycline inducible expression in fallopian tube cells	NA	High‐grade serous ovarian carcinoma	86
	Brca1	Brca1 loxP/loxP ;Trp53 loxP/loxP ;Rb1 loxP/loxP ;Nf1 loxP/loxP	Compound conditional	Exon 11 of Brca1	Ovgp1‐iCreER T2 ; Tamoxifen‐inducible expression in oviductal epithelial cells	NA	Serous tubal intraepithelial carcinomas, high‐grade serous ovarian carcinoma	87
	Brca2	Brca2 Δ11	Conditional	Exon 11 containing BRC repeats (RAD51‐binding motifs)	K14‐Cre; epithelial cells	NA	Development of mammary and skin tumors, reduced latency (181 d) in p53 fl/fl background	88
	Brca2	Brca2 G25R	Knock‐in	PALB2 interaction domain	NA	NA	B‐cell lymphoma and other tumors with long latency, increased tumor prevalence in p53 mutant background	89
	Brca2	Brca2 L2431P	Knock‐in	DSS1 interaction domain	NA	NA	Born in sub‐Mendelian ratio, live mice are radiation sensitive but no increased tumor development	90
	Brca2	Brca2 Δ27	Knockout	Exon 27 deletion, C‐terminal Rad51‐binding domain deletion	NA	NA	Viable in sub‐Mendelian ratio, develop range of tumors including sarcoma, carcinoma, adenoma, lymphoma (17‐19 mo latency), increased tumor incidences in p53 mutant background	91, 92
	Brca2	Brca2 Tr2014	Knockout	Truncation in aa 2014	NA	NA	Born in sub‐Mendelian ratio, adult live mice develop thymic lymphoma, show developmental defect	71
	Brca2	Brca2 tm1cam	Knockout	Truncation in exon 11	NA	NA	Born in sub‐Mendelian ratio, live mice develop thymic lymphoma, die within 12‐14 wk of age	72
	Brca2	Brca2 loxp/loxp ;Trp53 loxp/loxp ;Pten loxp/loxp	Compound conditional	Exon 11 of Brca2	Pax8‐rtTA; TetO‐Cre; Doxycycline‐inducible deletion in fallopian tube cells	NA	High‐grade serous ovarian cancer	86
	Palb2	Palb2 Δ2–3	Conditional	Exons 2 and 3	K14‐Cre; epithelial cells	NA	Mammary and skin tumors, latency reduced (192 d) in p53 fl/fl background	93
	Palb2 CC6/CC6	Knock‐in	Amino acids 24‐26 (LKK to AAA)		NA	Reduced male fertility, no gross developmental defect, MMC sensitivity	94

Abbreviations: GI, gastrointestinal; MSI, microsatellite instability; NA, not applicable.

The Brca1 and 2 mouse models have been very useful to study the mechanisms of tumor suppression and the role of different domains of the proteins in tumor development. Currently available models have shortcomings in that the heterozygous mice are not susceptible to development of tumors as do heterozygous humans and the tumors developed in these mice do not completely mimic human BRCA1/2 tumors. However, the models have been useful in preclinical studies of poly‐(ADP‐ribose)‐polymerase‐1 (PARP1) inhibitor or chemopreventive studies of progesterone antagonist (mifepristone) and RANKL (receptor activator of NF‐κB ligand) inhibitor. ⁷³ , ⁷⁴ , ⁷⁵ A preclinical study with Brca1‐deficient GEMMs showed that early introduction of a PARP inhibitor delays tumor development. ⁷⁶

6. CONCLUSIONS

Mouse models provide valuable resources for molecular analyses of precancerous and cancerous lesions in a defined genetic background, which is often difficult to accomplish by studying human population. The ability to capture de novo evolution of cancer due to certain mutations where immune cells are intact makes GEMMs attractive for immunoprevention studies which may provide useful insights into therapy response and resistance. Genetic homogeneity in mouse models allows the precise definition of the factors involved in the development of disease in a well‐defined genetic context. The different conventional and conditional knockout mice models presented here can clearly be useful for studying the mechanism of disease development and for cancer‐therapeutic and prevention studies. Long latency period and lower penetrance in many knockout models can make in vivo experiments complex and costly and thus hinder their use as a tool for large‐scale drug discovery. We anticipate that the current progress in genetic engineering technologies, especially with CRISPR‐Cas technology, will advance the rapid generation of multiallelic GEMMs and shorten the timelines for tumor development.

Next‐generation GEMMs with combinations of conditional knockouts for multigenic mutations in a specific organ or tissues in a time‐controlled fashion may more faithfully mimic the tumor development of patients with hereditary cancer syndromes. The development of these next‐generation GEMMs can further improve the effectiveness of GEMMs in predicting clinical efficacy and thus can potentially provide better management of hereditary cancer syndromes.

CONFLICT OF INTEREST STATEMENT

The authors have declared that no conflict of interest exists. Dr Robert H. Shoemaker is member of the editorial board.

DISCLAIMER

The opinions expressed by the authors are their own, and this material should not be interpreted as representing the official viewpoint of the U.S. Department of Health and Human Services, the National Institutes of Health, or the National Cancer Institute.

ETHICAL STATEMENT

Approval of the research protocol by an Institutional Reviewer Board. N/A.

Informed Consent. N/A.

Registry and the Registration No. of the study/trial. N/A.

Animal Studies. N/A.

Supporting information

Appendix S1:

Biswas K, Mohammed A, Sharan SK, Shoemaker RH. Genetically engineered mouse models for hereditary cancer syndromes. Cancer Sci. 2023;114:1800‐1815. doi: 10.1111/cas.15737