Diverse Landscape of Genetically Engineered Mouse Models: Genomic and Molecular Insights into Prostate Cancer

Abstract

Prostate cancer (PCa) is a significant health concern for men worldwide and is particularly prevalent in the United States. It is a complex disease presenting different molecular subtypes and varying degrees of aggressiveness. Transgenic/genetically engineered mouse models (GEMMs) greatly enhanced our understanding of the intricate molecular processes that underlie PCa progression and advancement and have offered valuable insights into potential therapeutic targets for this disease. The integration of whole-exome and whole-genome sequencing, along with expression profiling, has played a pivotal role in advancing GEMMs by facilitating the identification of genetic alterations driving PCa development. This review focuses on genetically modified mice classified into the first and second generations of PCa models. We summarize whether models created by manipulating the function of specific genes replicate the consequences of genomic alterations observed in human PCa, including early and later disease stages. We discuss cases where GEMMs did not fully exhibit the expected human PCa phenotypes and possible causes of the failure. Here, we summarize the comprehensive understanding, recent advances, and the strengths and limitations of the GEMMs in advancing our insights into PCa, offering genetic and molecular perspectives for developing novel GEMM models.

1. Introduction

Prostate cancer (PCa) is the second most frequently diagnosed non-cutaneous cancer and the fifth leading cause of cancer-related death in men globally. In the United States, it is a prevalent form of cancer in men, accounting for approximately 14% of all new cancer cases [1][2]. Radical prostatectomy remains a gold standard for treating localized PCa, and androgen deprivation therapy (ADT) is the backbone of systemic therapy for those who experience disease relapse after local therapy or for those with metastatic disease [3][4][5]. Poor prognoses persist in patients with advanced disease, attributed to factors like age, race, oxidative stress, inflammation, disease heterogeneity, drug resistance, and distant metastases [6][7]. As a result, the five-year disease-specific survival rate for metastatic disease stands at approximately 32.3% [1].

Creating genetically engineered mouse models (GEMMs) is crucial for studying diseases like cancers, allowing researchers to mimic human pathology [8][9]. During the 1980s, independent researchers unknowingly developed the first transgenic mice expressing dominant oncogenes (oncomice), highlighting concurrent progress in genetic engineering [10]. A breakthrough in this area was the creation of tissue-specific promoters and “Cre-lox” technology [11][12]. In cancer biology, there are three major types of genetic modification: "knock-in", involving the insertion of a transgene or modified allele; the “knock-out”, which entails the targeted inactivation of a specific gene or DNA segment to assess its function; and the “conditional and/or inducible knock-out”, which allows for spatially or temporally specific knockout [13]. Several strategies are utilized to integrate a transgene into the mouse genome to generate transgenic and gene knockout/knockin mouse models [14][15]. Major approach to introduce germline genetic modifications in a mouse include zygote route using pronuclear microinjection, and embryonic stem (ES) cell mediated gene -targeted transgene approach [16]. In the standard transgenic approach, the target transgene is injected directly into the pronucleus of murine embryos at the pronuclear stage [17][18]. The modified embryos are then transferred to surrogate mothers for gestation and development. Integration of the transgene into the genome typically occurs randomly. In recent years, exciting advancements in technology have been applied to the zygote route include the use of Sleeping Beauty transposons, ensuring reliable germline transgenesis[19]. On the other hand, in the gene-targeted transgenic approach, the transgene is introduced into ES cells [13][16]. The modified ES cells are then injected into blastocysts, which are subsequently implanted into surrogate mothers. This method allows for more precise targeting of the transgene to specific genomic loci. In addition to these, several other methods have been utilized in genetic research such as somatic cell nuclear transfer, retroviral -mediated transgenesis, homologous recombination in ES cells, sperm- and testicular-mediated gene transfer, Cre-lox mediated gene targeting, RNAi-mediated gene targeting, and Gene editing tool [15]. The advent of gene editing tools represents the most significant revolution in genetic research. This includes technologies like zinc-finger nucleases, transcription activator–like effector nucleases (TALENs), and clustered regularly interspersed short palindromic repeats (CRISPRs)/CRISPR-associated (Cas) systems [20][21][22]. The CRISPR/Cas9 system allows for precise insertion of transgenes into specific genomic locations in mice, ensuring high efficiency and control over integration sites, facilitating the generation of knock-in transgenic mice. Each of these strategies has its advantages and limitations, and the choice of method depends on factors such as the desired level of precision, efficiency, and experimental requirements.

Mouse modeling PCa is complex due to fundamental disparities in prostate biology and tumorigenesis among mice and humans. In comparison, one in six human males is at risk of developing PCa during their lifetime, whereas spontaneous PCa development is rare in mice, the translatability of findings [23]. Another challenge is that the mouse prostate is not a single organ but consists of four distinct lobes. In view of this, there are concerns regarding which mouse prostate lobe(s) resemble the human prostate the most. Further complicating the interpretation of experimental results is the variation in the stroma surrounding mouse and human prostate lobes [24][25]. This review aims to provide a comprehensive overview of the historical development and phenotypic description of the most widely utilized transgenic/genetically engineered mouse models of PCa that replicate the genomic variations that appear in human PCa. Moreover, the review also highlights the challenges/limitations of using mice models that are first- or second-generation derived. We have extensively discussed each model's conceptual grasp, current developments, and clinically pertinent issues and highlighted the strengths and challenges/limitations of first- or second-generation derived mice models to understand pathobiology and molecular aspects of PCa.

2. Comparative anatomy of mouse vs. Human prostate and morphological progression model of PCa

There are notable differences between the overall size, structure, and complexity of the glandular architecture of the human and mouse prostate [26]. The prostate gland in humans consists of three distinct histological zones: the peripheral (PZ), transitional (TZ), and central (CZ) zones. The PZ is the largest and most common site of PCa; the TZ surrounds the urethra and is a common site for developing benign prostatic hyperplasia (BPH), and the CZ surrounds the ejaculatory ducts (Figure 1A). On the other hand, the prostate gland in mice is divided into four lobes: the dorsal (DP), lateral (LP), anterior (AP), and ventral (VP) lobes, with the dorsal and lateral combined to form the dorsolateral lobe (DLP) (Figure 1A) [26][27]. Both human and mouse prostate epithelium contain luminal, basal, intermediate, and neuroendocrine cell populations (Figure 1 B, C); there is a notable difference in the arrangement of basal cells. The basal cell layer in human prostates appears continuously in histological sections, whereas in mouse prostates, basal cells connect through long cytoplasmic processes. Resultant a 1:1 ratio of basal to luminal cells in humans, while this ratio is ~1:7 in mice. The stromal tissue surrounding the epithelium varies between the two species [28].

Figure 1. Anatomy/ morphological structure of the murine and human prostate gland along with critical genetic/molecular alterations and schematic illustration of PCa evolution:

(A) The anatomical representation of the mouse and the human body and the location of the prostate gland. In mice, the prostate gland is located just below the bladder and surrounds the urethra. In humans, it is located in the pelvis, between the bladder and the rectum, and surrounds the urethra. (B) Diagram depicting the architecture of mouse and human prostate gland and adjacent structure. In mice, the prostate gland consists of four major lobes - the anterior, dorsal, lateral, and ventral lobes (left). In humans, the prostate gland is divided into three major prostatic zones - the peripheral zone, central zone, and transitional zone(right) (C) Representation of different cell types in the prostate. Histologically, it consists of secretory luminal, basal, intermediate, and neuroendocrine cells. The architecture of the prostate gland in both mice and humans consists of glandular epithelium embedded in the fibromuscular stroma. The prostatic epithelium is separated from the stroma by a basal lamina. It also comprises ducts and acini, which contain several cell types: smooth muscle cells, fibroblast, endothelial, and neurons. (D) A series of specific genetic and molecular alterations are involved in different stages of the PCa progression model, including PIN, localized adenocarcinoma, and invasive carcinoma. These alterations have been represented by deletion or inactivating mutation of the tumor-suppressor genes, overexpression of protooncogene, modulation of chemokines and pathway regulatory proteins, telomerase activation, and other molecular events as per disease stage (PIN, localized adenocarcinoma and metastasis). (E) Morphologic features of normal prostatic epithelium consisting of basal, luminal, and neuroendocrine cells. PCa progression occurs from either the luminal or basal cells of normal epithelium. At first, the development of premalignant lesions known as PIN has the potential to transform into low or high-grade carcinoma. Eventually, it becomes a metastatic disease spreading to the lymph nodes, bone, liver, and lungs via the circulation system. (F) The percentage bar graph presentation of human genomic alterations (Multiple alteration, fusion, truncating mutation, deletion, missense mutation, and amplification) concerning human metastasis and CRPC [9].

Notably, human and mouse tumors experience common biological processes driven by orthologous genetic events in their malignant evolution and exhibit noteworthy similarities in anatomy and morphology characteristics [29][30]. Regarding cellular populations, it has been observed that mouse distal and proximal luminal cells exhibit the closest similarity to human acinar and ductal populations, respectively. Nonetheless, here are some key points of comparison of the mouse and human prostate: (i) Lobes: Initially, VP in mice is often considered to be the most similar to the human peripheral zone, which is the region where most prostate cancers originate. Using single-cell analysis by Crowley et al., the similarity of cell composition between the mouse lateral prostate and the human peripheral zone was recently confirmed. The study showed that mouse distal and proximal luminal cells are most similar to human acinar and ductal populations. A periurethral population (PrU) that shares basal and luminal features is conserved between species [31]. (ii) Histology: Both mouse and human prostates comprise glandular tissue. They consist of epithelial cells arranged in glandular structures and are surrounded by stromal cells [32]. To account for differences and similarities in human and mouse prostate characteristics, investigators should use the ‘Bar Harbor Classification System’ when characterizing new GEMMs or undertaking experimental interventions that may impact the phenotype or natural history of lesion progression in existing models [21].

PCa is a multifaceted disease that develops, progresses, and advances through various genetic mutations, signaling pathway modulation, and epigenetic modifications that result in the dysregulation of vital cellular functions (Figure 1D). The progression model of PCa is represented (Figure 1E). Recent reports have summarized the human mutational and genomic alterations involved that lead to more advanced phenotypes often associated with castration-resistance prostate cancer (CRPC) and metastasis (Figure 1F) [33][9].

3. First-generation transgenic mouse models of PCa

3.1. Diverse promoters

A set of regulatory elements derived from different genes and regions have been effectively utilized to direct heterologous genes in prostatic epithelial cells of transgenic mice. These regulatory elements include (i) promoter/enhancer region from the rat C3(1) gene, (ii) androgen-sensitive rat probasin (PB), (iii) human prostate-specific antigen (PSA) gene and (iv) mouse mammary tumor virus (MMTV) long terminal repeat (LTR) [34][35][36][37]. The most commonly used promoter for developing transgenic mice lines is from the rat PB (PB), which is sufficient to drive androgen-responsive target expression in the prostatic epithelium [38][39][40]. PB promoter includes a minimal (−426 bp to +28 bp), PB DNA segment (also named PB), a larger (−11.5 kb to +28 bp) PB DNA segment (also named LPB), and a composite minimal PB DNA segment (also called ARR2PB) [35][39][41]. The ARR2PB-modified promoter was engineered by combining the minimal rat PB with an element of its enhancer region encompassing the two androgen-responsive elements (AREs) [35].

3.2. Viral oncogene-driven transgenic mouse models

First-generation transgenic mouse models have widely influenced PCa research and have provided a foundation for developing more advanced models (Figure 2A, B). The T antigen is an oncogenic viral gene in the genetic material of viruses like Simian virus 40 (SV40) and Polyoma virus [42] [43]. In the SV40 early region, two primary oncogenic proteins are prominent: the large tumor T antigen (Tag) and the small t antigen [44]. The large T antigen is responsible for cellular transformation via inactivating two essential tumor suppressor proteins, including p53 and RB [45][46][47]. The p53 transcription factor regulates the cell cycle and promotes apoptosis in cells under stress or damage. At the same time, RB controls the cell cycle by preventing cells from entering the S phase until they undergo proper growth control (Figure 2C). Similarly, the small t antigen interacts with the serine/threonine phosphatase PP2A and stimulates mitogen-activated protein kinase, which regulates several intracellular proteins implicated in various aspects of cellular transformation, including cell growth, proliferation, and survival thereby contributing to cancer development [48]. Therefore, the SV40 early region is considered an oncogenic "sledgehammer"[24].

Figure 2. Strategy for establishing a transgenic mouse model, traditional target vectors, and mode of the molecular action of SV-40 Tag:

(A) Standered transgenic approach [17][18]: Donor female mice were superovulated and allowed to mate overnight with stud male mice (1:1). Embryos were retrieved from the oviducts of donor females. Simultaneously, the preparation of the transgene construct and production of pseudopregnant recipient females were carried out. A transgene DNA fragment was injected into the male pronucleus of murine pronuclear stage embryos from the donor mice. Viable injected embryos were subsequently transferred to pseudopregnant recipient mice. Offspring were screened for the presence of the transgene through genotyping, and founders were identified. (B) Several promoters, including PB and non-PB systems, were used to generate transgenic mice. (C) Molecular mechanism of SV40-early region (large-T antigen) via targeting of tumor suppressor genes RB and P53. Specifically, the pRB-related protein family (pRB, p107, and p130/pRb2) generally binds to transcription factor E2F in early G1 of the cell cycle to prevent the expression of growth-stimulatory genes. The presence of SV40 T-ag disrupts this interaction, leading to the release of E2F and activation of these genes. In addition, in general, the role of P53 is to sense DNA damage and halt the cell cycle to allow for DNA repair or initiate apoptosis. The SV40 T-ag binds to P53 and blocks its function, allowing cells with genetic damage to survive and replicate.

The Male C3(1)-Tag mice are the first transgenic mouse line reported for studying PCa [34][49]. Various transgenic mouse lines have been created using different promoters, such as cryptdin 2/SV40 large-T/small-t, fetal gamma-globin/SV40 large-T/small-t (Gγ-globin-Tag), gp-91phox/SV40 large-T/small-t, MMTV/kgf, and others [50][51][52][53][54]. These lines exhibit distinct profiles of disease progression, as described in Table 1. A challenge with these transgenic mice is the risk of unintended off-target effects, resulting in transgene expression in other tissues due to the random integration of transgenes into the host genome.

Table 1.

Transgenic prostate cancer mouse model using non-probasin and prostate-specific probasin promoters.

Mouse Model	Features	References
C3 (1)-Tag	Develops prostatic epithelial hyperplasia in 12 weeks. Progression of prostate carcinoma with neuroendocrine differentiation within 28-44 weeks. Other infected tissues include the mammary gland, testis, and salivary gland. Metastatic sites infrequently to the lung.	[34] [155][156]
C3(1)/Py-MT	Develops prostatic lesions in both the ventral and dorsolateral lobe ranging from dysplasia to invasive carcinoma.	[157]
C3(1)/bcl-2	Proliferative lesions in the prostate epithelium in the ventral lobe, but unable to develop carcinoma and metastasis. Lesions were characterized by enhanced epithelial and stromal cell populations, which is similar to human BPH (benign prostatic hyperplasia). Transgene expression was also seen in the testis and uterus without phenotypic anomalies.	[158]
TRAMP (PB (−426 to +28bp) / SV40 early region)	Display mild to extensive hyperplasia with cribriform structures by 8-12 weeks, PIN within 18-24 weeks, and progression to carcinoma arises by 28 weeks. Develops poorly differentiated carcinomas with neuroendocrine phenotypes by 24 weeks. Metastasis frequently occurs in regional lymph nodes and lungs and less often in bone, kidney, and adrenal gland by 24-30 weeks. Castration-resistant phenotype.	[40][56] [61] [55][59][60]
LADY (PB (−11.5 to +28 kb) /SV40 Tag)	Develops prostatic epithelial hyperplasia by 10 weeks. Progressive lesions of high-grade dysplasia and poorly undifferentiated adenocarcinoma by 20 weeks. Commonly metastasize to regional lymph nodes, liver, and lung. Androgen-dependent tumor growth. Neuroendocrine phenotype.	[41][70][71]
Cryptdin 2/ SV 40 large-T/small-t	Develops rapidly progressive PCa with neuroendocrine features, followed by PIN to metastasis by 12 weeks, even without androgen stimulation. Commonly metastasize to lymph nodes, liver, lung, and bone by 16 weeks.	[50]
Fetal gamma globin/ SV 40 large-T/small-t (Gγ-globin-Tag)	Develops invasive carcinoma with neuroendocrine features by 16-20 weeks. Androgen-independent tumors are considered an accurate model for CRPC. Metastasize to the adrenal gland, lung, lymph nodes, bone, thymus, and intracapsular tissue of the neck and shoulders by 12-16 weeks.	[51][52]
gp-91phox/SV40 large-T/small-t	Develops invasive carcinoma with neuroendocrine features. Metastasize to surrounding tissue. Used for studying neuroectodermal malignancies.	[53]
MMTV/kgf	Develops prostatic epithelial hyperplasia in the ventral and dorsolateral lobe, seminal vesicles, and vas deferens. No invasive carcinoma, no metastasis. Transgene expression is noticed in the salivary and mammary glands of males.	[54]
MMTV/int2/fgf3	Develop enlarged ampullary glands, seminal vesicles, and ductus deferens but not the prostate.	[159]
MMTV/wap	Develop hyperplasia/dysplasia in the anterior gland epithelium with impaired mammary development. The transgene was expressed at elevated levels in organs with exocrine functions, including salivary and mammary glands, prostate, seminal vesicle, and coagulation gland. No invasive carcinoma, no metastasis.	[160]

Two major transgenic mouse lines have been created using distinct versions of the prostate-specific PB transcriptional regulatory elements. In the first mice model, the transgenic line carrying the rat PB gene was fused with the SV40 early region designated as the TRAMP (transgenic adenocarcinoma of the mouse prostate) line of mice. This model is the first autochthonous mouse model generated and characterized in 1996 [55][56]. This model exhibits various disease manifestations, and the median survival rate for these mice is 42 weeks. Most of these adenocarcinomas are confined to the DLP and VL epithelial cells, most analogous to the peripheral zone where most tumors originate in humans [57][58]. With a C57BL/6 genetic background, they histologically exhibit mild to extensive hyperplasia with cribriform structures by 8-12 weeks and later progress to carcinoma by 16-28 weeks [43]. By 24 to 30 weeks into the process, most of these mice have developed poorly differentiated carcinomas with neuroendocrine features and the ability to metastasize [37]. The castrating of TRAMP mice at 12 weeks of age showed no effect on primary tumor development or metastasis, as most (80%) of mice eventually developed poorly differentiated neuroendocrine carcinomas along with metastases by 24 weeks [60]. The incidence rate of metastasis in castrated mice was represented twice as compared to non-castrated TRAMP controls [60]. Remarkably, a mixed genetic background C57BL/6/FVBn exhibited enhanced tumor incidence in comparison to the parental strain and acquired bone metastasis at low frequency [8]. Hence, TRAMP model is considered as a reliable source of primary and metastatic tumors for in-depth analysis, aiming to elucidate the earliest molecular events in the onset, progression, and metastasis of PCa [61].

TRAMP mice model is referred to as a model to study NEPC characteristics; nevertheless, this mice model is recognized as effective for studying the mechanism involved in transforming “AR driven to non-AR driven” phenotypes of PCa [62][63][64]. In the TRAMP model, the disease advances progressively, displaying multifocal and heterogeneous characteristics [59]. Tang et al. utilized the TRAMP model to illustrate the tumor progression and the emergence of neuroendocrine carcinoma (NEC) in mice following chemotherapy (docetaxel) and/or hormone excision (castration). NEC was detected in both the castration-only and docetaxel-only treatment groups, with a notably higher incidence observed in the group receiving combined treatment (castration and docetaxel) [65]. Additionally, NEC tumors exhibited an elevated proliferative index, higher metastasis potential and drug-resistance [65]. Furthermore, Niu et al., employing prostate epithelial and stromal AR knockout or prostate epithelial-specific AR knockout TRAMP models, identified that the loss of AR in either mice led to the development of poorly differentiated primary tumors characterized by expanded intermediate cell populations [66]. In investigating the human genetic component of PCa susceptibility, a study explored how mouse strain background (FVB and B6) influences the TRAMP model. They observed that FVB-TRAMP mice had shorter lifespans due to increased neuroendocrine precursor lesions detected at 4 weeks compared to B6-TRAMP mice [67]. In addition, Gao et al., using the TRAMP model, investigated the involvement of the AKT1-β-catenin pathway in advanced PCa. The study revealed that Akt1 deficiency in mice inhibits oncogenic transformation, significantly reducing tumor size in TRAMP/Akt1^+/− mice and the complete absence of tumors in TRAMP/Akt1^−/− prostate. Conversely, late inhibition of Akt activity in 25-week-old tumor-bearing TRAMP mice promoted PCa metastasis to the lung [68]. Moreover, utilizing TRAMP model to examine the role of Znt7 (a Zinc transporter) revealed that Znt7-null TRAMP mice exhibited higher PIN and faster primary and metastatic prostate tumors [69]. Hence, the TRAMP mice model is considered the most suitable for investigating the potential role of gene and molecular mechanisms and studying agents that target hormone-refractory disease.

Another transgene line, the LPB-Tag (LADY) PCa mouse model, is pathologically related to the TRAMP model and was generated in 1998 [70]. Multiple LPB-Tag mouse lines have been created with a deletion in the early region to remove expression of the small Tag, ensuring consistent and elevated transgene expression in their prostates. Some contain a large fragment of PB upstream of the SV40 T-antigen, along with androgen and growth factor-responsive sequences. However, the chromosomal integration site still affects the expression level and prostate tumor growth rate. In this model, tumor growth is androgen-dependent since tumors relapse in mice castrated at 11 weeks, with no regrowth up to 40 weeks [70]. Remarkably, androgen treatment of castrated mice restores the epithelial cell versus stromal cell ratio and speeds up tumor growth. The most aggressive LPB-Tag line originated from 12T-7 and has numerous transgenes on two chromosomes. The offspring of this founder are classified into slow-growing 12T-7s and fast-growing 12T-7f lines. Both lines showed PIN with reduced invasive adenocarcinoma by 15-22 weeks. However, phenotypically, the 12T-7f line showed rapid prostate enlargement HGPINs and developed stromal hypercellularity [70]. In contrast to other mouse lines, 12T-10 has reflected slow prostate tumor growth and HGPINs without stromal hypercellularity, as seen in human HGPIN and PCa [71]. Hence, LADY mice are critical to analyzing the importance of neuroendocrine differences in the stepwise mechanism of PCa progression. The LADY model is also utilized to assess the effects of adjuvant therapy and chemotherapy agents. For instance, crossing LPB-Tag with VDR knockout mice is employed to investigate the role of VDR in prostate tumor initiation and progression [72].

The TRAMP/LADY models face limitations in oncogenesis research due to their reliance on a non-human-relevant viral oncogene for prostate-specific inactivation of pRb and p53. This leads to the development of "small cell" or NEPC tumors, which are rare in human PCa [73]. Despite these constraints, TRAMP and LPB-Tag models remain vital tools for comprehending molecular pathways in PCa and contribute significantly to preclinical chemoprevention approaches [74][75].

Table 1 describes the disease progression profile in these transgenic mice derived from the diverse promoter. Overall, T antigen-driven mouse models are widely used to study PCa biology because they can recapitulate certain characteristics of human PCa. However, these models present several drawbacks: (i) tumor development occurs at a much faster rate than typically observed in humans, limiting the study of PCa's natural history; (ii) these models often lack the histological heterogeneity seen in human PCa, a hallmark of the disease, (iii) T antigen-driven mouse models often rely on targeted genetic alterations in specific signaling pathways, restricting their ability to fully recapitulate the intricate genetic and epigenetic alterations occur in human PCa, (iv) the early expression of T antigen in these models affects prostate gland development starting from the first week of life [73][24][8][76].

4. Evolution of the Cre-loxP System: Advancements in transgenic mouse models

The Cre-loxP system was characterized by Sternberg’s laboratory from plaque-forming PI phage, which infects Escherichia coli [77]. Cre is a gene that encodes tyrosine site-specific recombinases that “causes recombination,” and loxP (locus of x-over, P1) is known as the “locus of phage crossing over.” In this way, Cre is a recombinase protein involved in site-specific genetic recombination (inversion, deletion, and translocation) in trans at loxP sites [77][12]. Cre was modified by fusing the mutant ligand binding domain of the human estrogen receptor (ER), activated by tamoxifen but not by estradiol, to accomplish more precise genetic studies and clinical applications. [78]. The benefits of conditional gene targeting are restricting Off-target effects and developmental defects. Moreover, it also allows for knocking down both alleles of interest without embryonic lethality. A strategy was employed to establish a transgenic mouse model (Figure 3A, B), resulting in the generation of various mouse lines outlined in Table 2.

Figure 3. Schematic representation of GEMM, common vectors, and mechanism of action of Cre-loxP system.

(A) Strategies for creating GEMM [16][13]: (1) blastocysts were extracted from mice, (2) embryonic stem (ES) cells were cultured, (3) a targeting vector containing DNA sequences homologous to the gene of interest is added in ES cells (4) and selected it for clonal positive selection (5) and targeted ES cells were expanded (6) Targeted ES cells were then injected into mouse blastocysts, (7) and implanted into a surrogate mother mouse, (8) developed the chimeric mouse. Screening of offspring on the basis of % presence of genetically engineered allele and this process repeated to create mice with different genetic alterations.

(B) Cre-lox system used to control site-specific recombination events in genomic DNA. When Cre recombinase encounters two LoxP sites, it catalyzes a recombination event between them. The result is a recombined locus in which the gene of interest is deleted, creating a “floxed” gene (left). For this purpose, several prostate-specific promoters have been developed, as represented (center). Two major systems, Cre-loxP and inducible Cre-loxP, were generated using these different vectors. The inducible Cre-loxP comprises Cre recombinase and a driver or promoter that controls when and where Cre is expressed, allowing for targeted gene expression manipulation (right). The principle of tamoxifen (T)-an inducible system of estrogen receptor fused to Cre (CreER) is: in the absence of T, expressed fusion protein, CreER, interacts with heat shock protein 90 (HSP90) and exists in the cytoplasm. (1) the interaction of HSP90 with CreER is disrupted after administration of T, (2) nuclear translocation of Cre happens with the interaction of ER and T, (3) the CreER recognizes the loxP sites in the nucleus, (4) and inactivates the gene X in a specific tissue.

Table 2.

Transgenic mouse lines expressing Cre recombinase/tamoxifen induced Cre recombinase in prostatic epithelial cells

Mouse Model	Recombinase	Recombinase activity	Ectopic Cre expression	Transcriptional regulatory elements	References
Pb-Cre	Cre - recombinase	VP > DP, AP, LP	Bladder, seminal vesicles	Rat PB (−426bp to+28bp)	[161]
Pb-Cre4	Cre - recombinase	LP > VP, DP, AP	Seminal vesicles, seminiferous tubules, prostatic Basal and stromal cell	ARR2PB	[162] [11]
ARR2PBi-Cre4	Cre - recombinase	LP, VP, DP, AP	Seminal vesicles, ductus deferens	ARR2PB	[163]
PSA-Cre	Cre - recombinase	Mature prostate lobes	-	PSA (6 kb)	[80] [79]
PSA-Cre-ERT2	Chimeric- Cre-recombinase; Cre-ERT2	Luminal cells; DLP, VP> AP	-	PSA (6 kb)	[164]
ARR2PB-Cre-ERT2	Chimeric- Cre-recombinase; Cre-ERT2	Luminal cells; AP, DLP> VP	-	ARR2PB	[165]
PB-MerCreMer	Chimeric- Cre-recombinase; MerCreMer	Luminal cells; AP, DLP> VP	-	PB genomic region	[166]
Nkx3.1Cre-ERT2	Chimeric- Cre-recombinase; Cre-ERT2	Luminal cells, Basal cells (10%)	-	Nkx3.1	[85]

The creation of the androgen-responsive ARR2PB (PB-Cre4) marks a substantial advancement in prostate-specific gene targeting, serving as a valuable tool for genetic-based PCa research [11]. Using the PSA-Cre promoter PSA^Cre;Nkx3.1^flox, and PSA^Cre;Pten^loxP/loxP mice were generated, showing PSA-Cre activity in all prostate lobes [79][80][81]. The Catnb^+/Δex3; MMTVcre mice were engineered to activate the Wnt/β-catenin pathway by deleting exon 3 in the β-catenin gene across various cell types. As a result, the stabilization of β-catenin in prostate epithelium led to hyperplasia and substantial transdifferentiation in to epidermal-like structures [82]. Similarly, the Pten^loxP/loxP;MMTV-cre mice were developed to selectively inactivate Pten in mouse tissue, resulting in the development of HGPIN that often advanced to focal invasive PCa [83]. However, these models have also been reported to develop off-target effects. FSP1(Fibroblast-specific protein 1) -Cre-driven gene deletion is reflected as a reliable strategy to explore fibroblast biology. The FSP1-Cre TGF^flox (Tgfbr2fspKO) mice model was generated to understand the role of transforming growth factor-beta (TGF-β) pathway and epithelial-mesenchymal connections by conditional inactivating the TGF- β type II receptor gene in fibroblasts [84]. The limitation of this promoter is gene recombination occurs in many tissues due to the universal presence of fibroblast throughout the body. Therefore, this model can be utilized to recognize the role of TGFβ in tissue differentiation but not a specific study for PCa. Next, tamoxifen induced Cre activity used to generate NKx3.1^CreERT2/+; R26R^YFP mice. Through genetic lineage-marking, it has been revealed that there are rare luminal cells within the prostate known as castration-resistant Nkx3.1-expressing cells (CARNs) exhibit bipotential properties and have self-renewal potential in vivo. Wang et al. demonstrated that these cells exhibit stem/progenitor properties within the androgen-deprived prostate epithelium during prostate regeneration, suggesting that Nkx3-1 is needed for stem cell maintenance [85]. In addition the targeted deletion of the Pten in CARNs (Nkx3.1^CreERT2/+;Pten^flox/flox mice), results in rapid onset of carcinoma subsequent to androgen-mediated regeneration [85]. Hence, this model has clinical relevance and could be targeted for CRPC.

5. Second generation genetically engineered mouse models of PCa: Integration of Human genomic alterations

GEMMs are sophisticated tools that have evolved from the single allelic transgenic model to generate complex multiallelic genetic alterations that more accurately mimic the genomic and molecular changes observed in human cancers, including PCa [9]. Table 3 presents an overview of GEMMs designed to mimic different stages of PCa and based on human genomic alteration. Table 4 represents the mice models based on AR, WNT, TGFβ, RAS/RAF/MAPK signaling dysregulation and epigenetic modifications.

Table 3.

GEMM based on human genomic alteration using different promoters.

Mouse Model	Features	Ref.
Pten-derived mouse model with multiple genetic mutations	Loss of function
Pten+/−	Develop dysplastic intestinal polyps, PIN, and neoplasia in multiple tissues. Most die within 8 months due to massive lymph-splenomegaly. The survivor PTEN+/− mice display PIN lesions by 10 months. No invasive and metastatic adenocarcinoma.	[89][88]
Ptenflox/flox(exon 4,5)/PBCre4	Develop HGPIN in 9 weeks, followed by carcinoma in 17-26 weeks. No metastasis up to 130 weeks.	[90]
Ptenflox/flox(exon 5)/PBCre4	Develop prostatic epithelial hyperplasia by 4 weeks, PIN in 6 weeks, and adenocarcinoma by 9-29 weeks. Metastasize to lung and lymph nodes by 12 weeks.	[95]
Pten+/−;p27−/−	Develop PCa at complete penetrance within 12 weeks from birth. No metastasis	[167]
Pten+/−; Nkx3.1−/−	Develop HGPIN by 6 months of age and ultimately progress to adenocarcinoma.	[117]
Pten+/−; Nkx3.1+/−p27+/− or Pten+/−; Nkx3.1+/−p27−/−	PCa progression is enhanced by the loss of one wildtype p27kip1 allele but inhibited by the loss of both alleles.	[97]
NPKEYFP mice (Nkx3.1CreERT2/+; Ptenflox/flox; KrasLSL-G12D/+; R26R-CAG LSL-EYFP/+	Tumor formation at 2- 3 months of age by tamoxifen administration and micro-metastasis at 5-8 months.	[102]
Pten−/+; Trp53−/− (RapidCap- model)	Metastasis to distant sites at greater than 50% penetrance by four months.	[105]
NPp53 mice (Ptenflox/flox; Trp53flox/flox; Nkx3.1CreERT2)	Under castrate, develops CRPC characteristics and, after treatment with abiraterone, displays NEPC features.	[109]
Ptenflox/flox: Rbflox/flox/ PBCre4 (DKO model)	Develops PIN and early invasive carcinoma within 12 weeks. Neuroendocrine features (synaptophysin-positive cells) appear by 20-25 weeks. ADT sensitive because castration extended the median survival of these mice from 38 to 48 weeks, but all mice eventually died from prostate cancer by 67 weeks. Median survival 38 weeks. Develops metastasis in the lymph nodes, liver, and lung.	[122]
Ptenflox/flox;Rbflox/flox;Trpflox/flox/ PBCre4 (TKO model)	Develops castration-resistant tumors, survival rate 16 weeks. ADT-resistance because castration of these mice at 10 weeks did not extend survival. These TKO tumors have reduced dependence on both AKT and AR signaling.	[122]
Hi-Myc-derived mouse model with multiple genetic alterations	Conditional Gain of function
Hi-Myc/ARR2PB; Hepsin	Invasive adenocarcinoma at 4.5 months. This model develops a higher-grade adenocarcinoma in 12- to 17-month-old mice. Tumors showed higher hepsin expression. The proposed mechanism is basement membrane degradation. No metastasis.	[168]
Hi-Myc/ARR2PB; IκBa+/−	Develops adenocarcinoma at 26 weeks. PCa progression to CRPC. The proposed mechanism is the activation of NF-κB signaling by regulating AR action.	[169]
Z-Myc-derived mouse model with multiple genetic alterations	Conditional loss of function
Z-Myc/PBcre4; PtenFl/Fl	HGPIN/cancer lesions and promote faster prostate tumorigenesis. The proposed mechanism is a bypass from senescence to apoptosis by repressing the p53 target gene p21Cip1.	[170]
Z-Myc/CMV-β-actin; Nkx3.1	HGPIN with microinvasion. MYC knockdown dysregulates shared NKX3.1/MYC target gene expression. The enhanced MYC transcriptional activity.	[171]
Z-Myc/PBcre4; Pten+/−;TrP53−/−	PIN and adenocarcinoma (10-24 weeks) with marked heterogeneity within the same prostate glands. It accelerates tumor formation and leads to lymph node metastases.	[172]
BMPC model (Hoxb13-Myc;Hoxb13-Cre PtenFl/Fl)	Develops lethal adenocarcinoma with distant metastases. BMPC cancers are deficient in neuroendocrine/ or sarcomatoid differentiation. Induce genomic instability and aggressive prostate cancer similar to human disease at the histologic and genomic levels.	[98]
Other models	Conditional modulation of gene expression
Trp53−/−; Rb−/−	PIN lesions at 8 weeks. Poorly differentiated neuroendocrine features by ~32 week Tumors are ADT-resistant de novo because castration does not extend survival.	[73]
Trp53R270H/+ Nkx3.1Cre	PIN as early as 5 weeks, well-developed neoplastic foci as early as 4 months. Displays invasive adenocarcinoma with epithelial-mesenchymal transition (EMT) phenotype.	[108]
Tmprss2-Erg fusion/Pten−/+	Acceleration of HGPIN and progression to prostatic adenocarcinoma at 6 months. The proposed mechanism is by modulating AR signaling and checkpoint genes.	[135]
Spop/Pten−/+	Develops early neoplastic lesions (HGPIN) with striking nuclear atypia and invasive poorly differentiated carcinoma. Activates both PI3K/mTOR and AR signaling.	[138]
PBCre4/Foxa1loxp/loxp	Develops hyperplasia with an extensive cribriform pattern but does not expand toward more advanced phenotypes. Alterations in localization of p63 positive basal cells. Following castration, the conditional knockout Foxa1 retained its androgen sensitivity, and Foxa1-positive cells appeared to be more castration sensitive.	[173]

Table 4.

GEMMs through alteration of molecular signaling pathways:

Molecular signaling pathways	Features	Reference
AR	AR transgene expression in the prostate epithelium develops PIN.	[174]
	The mutant AR (AR-E231G) expressed under the control of the PB promoter facilitates the PIN. This condition subsequently progressed to invasive and metastatic disease.	[175]
	The PB-driven ARv567es transgenic mouse model displays epithelial hyperplasia at 16 weeks and develops invasive adenocarcinoma by 1 year of age.	[176]
	The transgenic line (R26hAR(loxP): Osr1-Cre+) exhibits conditionally activated AR. Transgene expression is seen in both prostatic luminal and basal epithelial cells and is resistant to castration.	[177]
Wnt/ β-catenin	Conditional loss of function of adenomatous polyposis coli (Apc) develops HGPIN and adenocarcinoma depending on Cre-driver.	[178]
	Conditional activation of β-catenin represents PIN phenotypes, and cancer progression is achieved by collaboration with the loss of function of Pten	[179] [180] [181]
	Synergism of activated K-ras (K-ras+/V12) and dominant stabilized β-catenin (Catnb+/lox(ex3)) with PbCre promoter induces rapid progression of prostate tumorigenesis to invasive carcinoma.	[182]
	sδ-catenin (gene designation: CTNND2) mutations promote tumor development in mouse prostate with ARR2PB promoter	[183]
	Combining Wnt-signaling and Tag expression in the mouse prostate (LPB-Tag/dominant active mice) display neuroendocrine differentiation (NED), but NE cancer does not develop.	[184]
TGF- β	Model in PCa to study stromal interaction.	[84]
	PbCre4 promoter-dependent conditional loss of Smad4 used for TGF-β activation. Combined conditional loss of Pten and Smad4 develop aggressive adenocarcinoma, characterized by a short latency period and distant metastasis with low penetrance.	[100]
	In addition, conditional loss of Pten, Smad4 and Trp53, along with telomerase deficiency, represent invasive adenocarcinoma with extensive metastasis (bone metastasis).	[110]
	COUP-TFII (NR2F2) inhibits Smad4-dependent transcription, and overexpression of COUP-TFII in the mouse prostate epithelium cooperates with Pten deletion develops an aggressive metastasis-prone tumour.	[185]
Ras/Raf/MAPK	Conditional gain of expression of mutated gene Kras G12D with Pten-based background in the prostate of mice represents invasive adenocarcinoma with and macrometastasis with 100% penetrance.	[186]
	An activating mutation of BrafV600E induced GEM to develop prostate gland hyperplasia with rapidly growing invasive adenocarcinoma.	[187]
	Conditional gain of expression of NCoA2 (nuclear receptor coactivator 2) with Pten loss of function background shows invasive adenocarcinoma with distant metastasis.	[188]
Epigenetic regulator	Histone methyltransferase WHSC1 overexpression in prostate epithelium, combined with Pten deletion, resulted in a metastasis-prone tumor in a Pten-null murine PCa model.	[189]

5.1. GEMMs based on tumor suppressor genes

5.1.1. Pten-derived mouse model with multiple genetic mutations

PTEN (phosphatase and tensin homolog deleted from chromosome 10), known as Mmac1, is a crucial tumor suppressor gene, and its loss of function is associated with ~ 35% of primary PCa and ~63% of advanced/ metastasis [86][87]. The first Pten knockout mouse was developed by creating a null mutation in 1998; one more group in 1999 also created a Pten knockout mouse by deleting the distinct portions of the Pten gene that showed similar inactivating activity [88][89]. These null models were embryonically lethal, and no homozygous knockouts were alive [89]. The various models of Pten-based GEMM have been developed, which differ based on the severity of disease phenotypes, prevalence, latency, and biology. The differences in severity depend on the timing of gene recombination comparative to prostate differentiation, contextually to specific cell types targeted for recombination, selection of precise Cre-driver, specific deleted region of Pten exon, mouse strain background, and variance in phenotypic-pathological interpretation [90][91]. To assess the impact of varying PTEN levels on cancer progression through homologous recombination, a series of hypomorphic Pten mouse mutants ( Pten^hy/+, Pten^+/−, and Pten^hy/−) with decreasing Pten activity was created [90]. It has been observed that the Pten dosage dictates the PCa progression and regulates critical downstream targets, including Akt, p27(Kip1), mTOR, and FOXO3 [90]. A recent study highlighted the importance of the dynamic interplay between mTOR and AMPK activities in metastatic colonization by circulating PCa cells, emphasizing the concurrent targeting of both AMPK and mTOR [92]. Bertram et al. reported that Pten inactivation is adequate for acquiring androgen independence progression [93]. A recent study investigated the impact of age on PCa using a controlled Pten-null mouse model; both aged and non-aged adult mice developed prostatic epithelial hyperplasia within 4 weeks, PIN within 8 weeks, and some PINs progressed to invasive adenocarcinoma between 8 to 16 weeks post-Pten ablation. Notably, aged mice showed accelerated PI3K/AKT/mTOR signaling and increased onset and progression of PCa compared to young mice [94].

The conditional Pten deletion in the prostatic epithelium showed progression towards hyperplasia after 4 weeks, PIN in 6 weeks, and full adenocarcinoma (100% penetrance) at 9-29 weeks [95]. Histopathologically, the Pten null tumor originated from secretive epithelial, a source similar to most human prostate tumors [95]. Notably, the invasive Pten-null PCa cells did respond to androgen ablation and showed increased apoptosis, mirroring human PCa. Even though the survival of these cells was androgen-sensitive, their proliferation remained unaffected. A global evaluation of molecular changes in Pten-null mice (homozygous deletion) identified the “signature” gene expression changes similar to those relevant for human PCa and other human cancers with metastatic potential. This model provided an exclusive tool for investigating the molecular process of PCa metastasis and developing new targeted therapy [95]. The Pten mouse model has some limitations in its histological representation, such as the development of squamous differentiation, sarcomatoid carcinoma, and mucinous metaplasia in some instances [96].

Interestingly, triple mutant mice (Nkx3.1^+/− or ^−/−; Pten^+/−; p27^+/−) that were heterozygous for a p27^kip1 null allele exhibited accelerated prostate carcinogenesis, which is contrary to the findings observed in homozygous null for p27Kip1 (Nkx3.1^+/− or ^−/−; Pten^+/−; p27^−/−) mice, which showed suppression of PCa progression [97]. Expression profiling revealed that compound p27kip1 heterozygotes exhibited higher Cyclin D1 levels, while these levels were downregulated in compound p27^kip1 homozygous mutants [97]. The loss of Pten with other regulatory proteins has also been investigated; all contribute to PCa progression and metastatic phenotypes. These includes MYC activation (Hoxb13-MYC;Hoxb13-Cre;Pten^Fl/Fl), participation of ETV4 via PI3K and RAS signaling and the functional significance of SMAD4 via TGFβ/BMP–SMAD4 signaling (Pten^loxp/loxp;Smad4^loxp/loxp;PB-Cre4) [98][99][100]. Furthermore, a study indicated that the transgenic expression of a constitutively activated p110β (Pi3kcb) allele in the prostate leads to PIN formation. The resulting lesions resemble those caused by PTEN loss or Akt activation but exhibit clear distinctions [101].

Recently, a new NPK^EYFP mice model (Nkx3.1^CreERT2/+; Pten^flox/flox; Kras^LSL-G12D/+; R26R-CAG ^LSL-EYFP/+) was developed for detailed biological and molecular characterization of lethal PCa related to bone metastasis and castration-resistance [102]. These mice use the inducible Cre guided by NKX3.1 promoter to accomplish spatiotemporal regulation of gene recombination (Pten^flox/flox; Kras^LSL-G12D/+), particularly in prostatic luminal cells. Visualization of tumor and bone metastasis is made possible by activating the fluorescent reporter allele (R26R-CAG ^LSL-EYFP/+). Through tamoxifen administration, NPK^EYFP mice were induced to tumor formation at 2-3 months and monitored for micro-metastasis at 5-8 months. This model developed PCa with high penetrance for metastasis and is capable of homing to the bone. The median survival of NPKEYFP mice (4.7 months) was significantly higher than control NPK (3.1 months) mice. Sequencing data analysis revealed a link between MYC/RAS co-activation and PCa metastasis, underscoring the importance of the META-16 gene profile in adverse outcomes and treatment response [102].

5.1.2. TrP53-derived mouse model with multiple genetic alterations

TRP53 mutated or deleted in >50% of aggressive PCa [103]. The alteration of P53 expression either by losing one copy of the Trp53 gene or by a mutation in transgenic mice PCa model presented HGPIN lesions (grades III-IV) by 52 weeks with decreased apoptotic rate [104]. The germline or conditional loss of function of Trp53 based GEMMs model represented modest PCa phenotypes, with only PIN lesions not transformed towards adenocarcinoma.

To evaluate the therapeutic potential for metastatic PCa, the RapidCaP mouse model was designed to utilize a surgical technique to introduce lentiviral transgenes into the prostate. This model is established with the combined loss of function of Trp53 and Pten genes and displays metastasis to distant sites (> 50% penetrance) within 4 months [105]. Prostate-specific deletion of Pten and Trp53 by incorporating a Cre-activatable luciferase reporter resulted in uniformly lethal disease by 25 weeks [106]. The functional consequence of Trp53 alteration in conjunction with Pten loss in the prostate epithelium is cell lineage plasticity [107]. Another study revealed that prostate-specific Trp53 mutant Trp53R270H expression, i.e., similar to the human hotspot mutant R273H, indicated a moderate PIN phenotype [108]. This feature was augmented with a combination of loss of function of Nkx3.1^+/−. Furthermore, the related loss of Trp53 and Pten with Nkx3.1, Pten ^flox/flox; Trp53^flox/flox; Nkx3.1^CreERT2 (NPp53 mice) following castration encouraged CRPC with neuroendocrine differentiation phenotypes [109]. These NPp53 mice are highly aggressive and resistant to anti-androgen treatment, and notably, anti-androgen therapeutic approaches aggravate disease progression [109]. Preclinical analyses revealed that abiraterone treatment caused modest but significant suppression of NP CRPC. At the same time, NPp53 CRPC models were resistant to abiraterone, as apparent from their histopathology and absence of reduced cellular proliferation or tumor volume. The NPp53 mouse model displays accelerated tumor progression, with features similar to treatment-related CRPC with neuroendocrine differentiation as seen in humans. Therefore, these models are not the best choice for studying the efficacy of abiraterone or other AR-targeted therapies in CRPC [109]. Moreover, it has been speculated that telomerase reactivation following telomere dysfunction generates prostate tumors with bone metastases. Study reported that the inducible telomerase reverse transcriptase (mTert) allele crossed over with Pten null and Trp53 developed aggressive cancers with rearranged genomes and new tumor biological features [110].

5.1.3. Nkx3.1-derived mouse model with multiple genetic mutations

The allelic deletion of NKX3.1 has been reported in ~80% of prostatic neoplasia and is considered a gatekeeper suppressor gene [111][112]. The first Nkx3.1 knockout mouse was established in 1999 to examine its physiological effects, and numerous phenotypical changes were observed in the prostate and seminal vesicles [113]. Loss of Nkx3.1 (conventional and conditional) promoted initiation of dysplasia and hyperplasia in an age-dependent manner, but no overt tumors were observed [79][112]. The targeted interruption of Nkx3.1 in mice showed morphogenetic defects in the salivary gland, suggesting its role in developing other cell types [114]. Bowen et. al investigated that the absence of NKX3.1 expression is highly correlated with hormone-refractory disease and advanced tumor stage in PCa [115]. In another study, aged Nkx3.1 mutant mice exhibited histopathological abnormalities resembling PIN and suggested that the loss of Nkx3.1 function is a pivotal event in the initiation of prostate cancer and serves as a model for early-stage disease [116].

Notably, Nkx3.1; Pten double mutant mice showed an augmented incidence of HGPIN, invasive adenocarcinoma, and lymph node metastasis, highly similar to the early stage of human PCa [117]. Loss of Nkx3.1 coordinates with Pten loss for PCa progression, and this cooperativity is mediated by synergistic activation of AKT /protein kinase B [117]. Another finding affirmed the function of PTEN as a phosphatase for NKX3.1, opposing phosphorylation at NKX3.1(S185) to preserve its stability and extend half-life. In gene-targeted mice, Pten loss significantly reduced Nkx3.1 expression, promoting prostate epithelial cell proliferation [118].

5.1.4. Rb-derived mouse model with multiple genetic alterations

Retinoblastoma (RB), a tumor suppressor gene at chromosomal loci 13q, exhibits a specific mutation linked to early events in PCa [119][120]. Functional RB family proteins (Rb/p107/p130) inactivation via conditional deletion of the RB gene in prostate epithelium stimulates epithelial proliferation and apoptosis. This model produced PIN and adenocarcinomas with no indication of neuroendocrine tumors [120]. The conditional loss of Rb function in the prostate reflected a relatively modest phenotype, but collaboration with other genomic alterations accelerated PCa progression [121][73]. In particular, a Trp53^PE−/−: Rb^PE−/− knockout mice model developed PIN lesions as early as 8 weeks and a poorly differentiated phenotype with neuroendocrine features by 32 weeks. Rb and TrP53 loss was observed to suppress the epigenetic reprogramming factors (EZH2 and SOX2), creating a stem-cell-like epigenetic environment permissive to lineage plasticity [73].

Next, DKO mice (PBCre4: Pten^f/f: Rb^f/f) had short latency, established PIN lesions by 12 weeks, had distant metastasis, and displayed shorter survival (38 weeks) than single gene mutant models (48 weeks). In DKO mice, Rb loss acted as an enhancer of lineage plasticity stimulated by Pten loss. The DKO model, sensitive to ADT, exhibits reduced AR expression, spontaneous Trp53 mutations (V173M, R282Q), and tumors relapsed following castration [122]. TKO mice (PBCre4: Pten^f/f: Rb^f/f: Trp53 ^f/f) had a concise survival rate of 16 weeks and reflected aggressive phenotypes. Despite androgen deprivation, tumor biology, and survival were unaffected, confirming the insensitivity of these tumors to castration. This model unveiled that Rb1 loss facilitates lineage plasticity and metastasis initiated by Pten mutation, while additional Trp53 loss results in resistance to antiandrogen therapy [122].

5.2. GEMMs based on activation of gene

5.2.1. Myc-derived mouse model

The MYC overexpression in the prostate depends on the level of MYC expression as high (Hi) or low (Lo) GEMMs have been developed [123]. The PB/ Lo-Myc model comprised minimal PB promoter expressed a low level of MYC, and PIN lesions were perceived at 10 weeks, showing progressive to invasive adenocarcinoma by 10-12 months. In comparison, the ARR2PB / Hi-Myc model consists of a reconstructed PB promoter expressing a high level of MYC, and the PIN lesions were noticed at 2 weeks, showing prostate intraepithelial neoplasia to invasive adenocarcinoma by 3-6 months. Although PB/ Lo-Myc and ARR2PB / Hi-Myc mice showed similar phenotypes at pathological terms, the difference is only the duration of pathological changes and their androgen responsiveness [123][124]. Furthermore, Hi/Myc mice progressed towards CRPC following androgen deprivation [123]. A characteristic feature of Myc-based models is that they do not exhibit sarcomatoid or neuroendocrine differentiation histopathology [98]. Hence, this model is pronounced as an accurate representative model of human adenocarcinoma. Moreover, several Myc based GEMM cooperativity with other genes have been developed, as discussed in Table 3.

Particularly, the BMPC model (Hoxb13-Myc; Hoxb13-Cre; Pten^flox/flox) had a higher level of MYC and tended to develop invasive carcinoma with distant metastasis to lung, liver, and bone rarely. Tumors of BMPC mice displayed low AR and NKX3.1 but did not exhibit sarcomatoid characteristics prevalent in other advanced PCa GEMMs [98]. Dysregulation of N-MYC cooperating with myristoylated AKT1 led to the development of adenocarcinoma and NEPC, presenting phenotypic and molecular traits attributed to aggressive, late-stage human PCa [125]. Pharmacologic disruption of N-MYC expression through Aurora A kinase inhibition (AURKA inhibitor CD532) reduced the tumor burden [125].

The GEMM (T2-Cre^+/+; Pten^f/+; LSL-MycN^+/+) carried a CAG-driven lox-stop-lox human N-Myc gene incorporated into the ROSA26 (LSL-N-Myc) locus and a Tmprss2-driven tamoxifen-activated Cre recombinase (T2-Cre) that facilitates its expression in luminal prostate cells [126]. This model has also been engineered to harbor a Pten conditional knockout allele. In this model, tamoxifen-induced Cre activation resulted in N-MYC overexpression in the prostate, persisting from 5 weeks to 3 months post-induction, leading to HGPIN at 3 and 6 months. Subsequently, a substantial, invasive prostate tumor emerged at 9 months, featuring AR-positive adenocarcinoma, poorly differentiated carcinoma foci, and divergent differentiation [126].

5.2.2. MPAKT transgenic model

The tumor phenotype in Pten^−/−embryonic stem cells depends on AKT activation [127]. The MPAKT (murine prostate-restricted AKT kinase transgenic mice) model is generated by introducing a plasmid insert containing a PB promoter, a myristoylated sequence (myr), and a hemagglutinin (HA) epitope-tagged sequence, and human AKT1in the form of rPB-myr-HA-AKT1 insert into fertilized oocytes pronuclei. This resulted in the expression of the myristoylated AKT1 oncogene, specifically in the prostate epithelium. The MPAKT mice developed PIN lesions in the ventral prostate, accompanied by significant bladder obstruction, but without any signs of metastasis [128]. At the molecular level, the MPAKT mice represented activated AKT1 spatially confined to the prostate and showed activation of the p70S6K pathway. This model is valuable for examining the role of AKT/ PKB in prostate epithelial cell transformation and discovering biomarkers relevant to human PCa [95].

5.3. GEMMs based on genetic rearrangements, mutation, and alteration:

5.3.1. TMPRSS2-ERG fusion gene mouse model:

E26 transformation-specific (ETS) gene fusion and overexpression of ETS factors ERG, FLI1, ETV1, ETV4, and ETV5 have been reported in PCa [129]. Activation of ERG is associated with PCa progression in both early- and late stages [130][131]. Expression of the TMPRSS2:ERG gene fusion regulates bone markers and promotes the osteoblastic phenotype of PCa bone metastases [132]. Inducing targeted expression of the ERG in luminal prostate epithelial cells of transgenic mice reflected the initiation of prostate neoplasia, characterized by the development of focal precancerous PIN. Notably, like human cancers, luminal epithelial cells in these PIN lesions reduce the number of basal epithelial cells and make direct contact with the stromal cell compartment. Hence, this study indicates that ERG plays a vital role in the transformation of prostate epithelium and is considered a promising target for early therapeutic intervention [133]. Notably, PB-driven ERG, ARR2PB-driven ETV1, or ERG over-expression resulted in only PIN and no other phenotypes over the mouse lifetime [134][135]. In addition, ARR2PB-driven TMPRSS2-ERG exhibited no histological phenotype for up to 60 weeks [136]. The concomitant TMPRSS2-ERG expression and PTEN heterozygous background transgenic mice promoted the marked acceleration of HGPIN and invasive prostatic adenocarcinoma at 6 months of age [135].

5.3.2. SPOP and other early alteration-related mouse models:

Genomic instability/ impaired genome maintenance resulting from point mutations in E3 ubiquitin ligase SPOP (Speckle-type POZ protein) is a typical mechanism driving prostate tumorigenesis and occurs in ~10% of PCa [137]. SPOP-mutant PCa represents genetically and biologically distinct subsets that respond to AR signaling inhibition [7]. A GEMM with prostate-specific conditional expression of mutant Spop (Spop^F133V) showed a modest phenotype. The setting of Pten loss with Spop mutation reflected dramatically altered phenotypes such as early neoplastic lesions (HGPIN) with dramatic nuclear atypia and invasive poorly differentiated carcinoma [138].

The SPOP mutations and CHD1 (chromatin remodeler chromodomain helicase DNA-binding protein 1) homozygous deletions frequently co-occur in PCa patients [139][140]. The GEMM PB-Cre⁺; Chd1^L/L mice led to the development of PIN but did not develop PCa. Mechanistically, this model showed changes in DDR (DNA damage repair) by modulating the choice between HR (homologous recombination) and NHEJ (non-homologous end joining) DNA double-strand breaks (DSBs) repair. Remarkably, loss of CHD1 reduces HR-mediated DSB repair and enhances error-prone NHEJ activity [139]. FOXA1 (forkhead box) alteration in PCa was mutually exclusive with the alteration of SPOP and ETS fusion [141]. The impact of prostate-specific FOXA1 deletion in the epithelium of adult mice indicated progressive florid hyperplasia with an extensive cribriform pattern but did not progress toward more advanced phenotypes [142].

Conclusion & Future Insights

Overall, mouse models have proven valuable tools for PCa research but have several limitations, as discussed. An ideal mouse model should adhere to the following principles: (i) it should encompass all stages of PCa initiation and progression, exhibiting easily identifiable features that mirror both the histopathological and molecular aspects of human PCa; (ii) at early stages, the model should demonstrate AR biology and exhibit a slow progression rate; additionally, it should possess the ability to metastasize to the same sites as human PCa, while maintaining an intact immune system. The emerging concept of personalized mouse models, Co-clinical Trials, has been introduced to overcome these pitfalls. These trials enable the real-time integration of the murine and human tumor data. In Co-clinical Trials, GEMMs are utilized to guide therapy in an ongoing human patient trial toward future clinical management of the patient's tumor [143].

GEMMs are considered indispensable for preclinical research and superior to cancer cell inoculation models. One unique feature of GEMMs is the ability to develop de novo tumors in a natural immune-proficient microenvironment. In addition, advanced GEMM-derived tumors closely resemble the histopathological and molecular characteristics of their human counterparts, exhibiting genetic diversity and progressing spontaneously toward metastatic conditions [144]. Furthermore, GEMMs of PCa represent potent tools for investigating the mechanisms underlying disease progression and the response to therapy. A recent study detailed the development, examination, and validation of a platform designed to immobilize and precisely target tumors in mice using stereotactic ablative radiation therapy (SART). This approach has been observed to modify the tumor stroma and immune environment, improving survival outcomes in GEMMs of primary PCa and exhibiting synergy with androgen deprivation [145].

Nevertheless, despite the SART application, complete pathologic responses were not attained in GEMMs. Consequently, investigating the mechanisms of radiation resistance in GEMMs of PCa presents an intriguing area for further exploration. The onset, advancement, and development of CRPC entail intricate interactions between tumor cells and the host immune system [146]. GEMMs offer a unique prospect to investigate the inherent factors influencing tumor response and assess the impact of extrinsic determinants. It has been reported that GEMMs of PCa have differing quantities of tumor-infiltrating lymphocytes characterized by notable spatial heterogeneity [145].

Using these mouse models has led to the development of new imaging techniques that can detect bone metastases at an earlier stage and tumor microenvironment mechanisms for bone metastatic disease progression, potentially allowing for more effective treatment [102][147]. Human PCa represents a heterogeneous diversity at inter- and intra-tumor and inter-patient levels, which significantly influences the prognosis, choice of therapy, recurrence, and emergence of treatment resistance in PCa [148]. In treatment-naive primary tumors, there is pre-existing AR heterogeneity, which is magnified during AR signaling inhibitors (ARSIs) treatment and notably pronounced in CRPC [149]. After extensive research spanning several decades to create GEMMs with molecular changes akin to human PCa, no mouse model has fully reflected the heterogeneity of human tumors or accurately recapitulated the observed phenotype in humans due to genetic, evolutionary, species-specific, and clinical heterogeneity [150][151][152].

Initially, basal cells were identified as the originating cells for PCa, but emerging evidence indicates that luminal cells are now considered the preferred cell of origin for PCa [153]. The emergence of single-cell RNA-sequencing technology has offered new insights into prostate epithelium heterogeneity, which is conserved between mouse and human [154]. Therefore, basal and luminal specific promoters can be valuable in identifying drivers of PCa progression precisely to each compartment.

Nevertheless, many genes remain untested, and ongoing research is necessary to uncover new genes and improve the complexity of mouse models to reflect human PCa better. The field of GEMMs for PCa is entering a new phase known as next-generation GEMMs, characterized by applying novel tools and ideas to enhance our understanding of lethal disease biology. These advancements have the potential to lead to the foundation for more accurate and relevant models for studying PCa and developing new therapeutics.