Modelling human prostate cancer: Rat models

ABSTRACT Prostate cancer is the second most common cancer in men, affecting approximately 1.1 million men worldwide. In this way, the study of prostate cancer biopathology and the study of new potential therapies is of paramount importance. Several rat models were developed over the years to study prostate cancer, namely spontaneous models, chemically‐induced models, implantation of cancer cell lines and genetically‐engineered models. This manuscript aimed to provide the readers with an overview of the rat models of prostate cancer, highlighting their advantages and disadvantages, as well as, their applications. Graphical abstract: Figure. No caption available.


Rat as a model of prostate carcinogenesis
Despite many research projects in the field of prostate cancer are carried out using cells lines (in vitro studies), which allow the understanding of biological aspects related to the development of this disease, they fail to mimic the complex cellular interaction that occurs in tumor microenvironment. To overcome this limitation, researchers employed their efforts for several years on the development of animal models to study this disease.
In 1937, Moore and Melchionna were the first ones to induce prostate carcinoma in the anterior lobe of the White rat prostate through the direct injection of 1:2 benzpyrene into prostate [12]. Metastasis development was not reported in this model, being considered a model limitation. Some years later, in 1945, Dunning and colleagues, induced the development of metastasizing prostate carcinomas in albino Fisher 344 rats and black agouti Irish AxC 9935 rats [13], through the implantation of methylcholanthrene crystal into prostate. Since then, several experimental researches were conducted to discover chemical carcinogenesis with tropism for prostate tissue [14]. Highlight the N-nitrosobis (2-oxopropyl) amine (BOP) discovered in 1981 by Pour [15], the N-Methyl-N-nitrosourea (MNU) discovered in 1986 by Pollard [9], the 3,2'dimethyl-4-aminobiphenyl (DMAB) discovered in 1986 by Katayana [16] and the 2amino-1-methyl-6-phenylimidazol[4,5-b]pyridine (PhiP) discovered in 1997 by Shirai [17]. An adequate rat model of prostate carcinogenesis should develop androgen-sensitive adenocarcinomas in a short period of time, mimic the physiology and characteristics of man tumors, and the tumors must metastasize, preferably to bones. The tumor development should be from dorsal, lateral or anterior prostate lobes that are human homologue [14].

Rat models of prostate carcinogenesis
Several animal models are available for the study of prostate cancer: spontaneous tumors, chemically or hormonally-induced, implantation of cancer cells and genetically engineered animals [2,32].

Spontaneous tumors
The first report of prostate spontaneous tumor was in 1963 by Dr. W.F. Dunning, who detected a single adenocarcinoma without metastasis in a 22 months-old Copenhagen rat [33][34][35]. Later, in 1973, Pollard reported spontaneous prostatic carcinomas in dorso-lateral and anterior prostate lobes, in a germfree Wistar rats with 26 months of age [34]. In 1957, spontaneous tumors of the ventral prostate were reported in [34][35][36][37] months-old AxC rats [33]. Aging male, on average 24 months, ACI/Seg rats demonstrated high susceptibility to develop spontaneous ventral prostate adenocarcinomas [36]. Since a long latency period is needed for tumor development (around 2-3 years) and tumor incidence is low, these spontaneous tumors are not advantageous as animal models [32].

Induced tumors
The prostate tumors development may be easily induced through the administration of chemical carcinogens alone or by the combination of chemical carcinogens and hormones [2].

Chemically-induced prostate cancer
Up to now, four chemical compounds are recognized for the induction of prostate carcinomas development in rats: BOP, MNU, DMAB and PhiP [2].

N-nitrosobis (2-oxopropyl) amine (BOP)
The BOP belongs to the family of nitrosamines and it acts as a potent carcinogen in different species and organs, like hamsters (pancreatic tumors) [37,38] and guinea pigs (biliary and hepatic neoplasms) [39]. When ingested by the animals, this compound is converted into carbon dioxide becoming a DNA methylating agent. BOP is also converted in two nitrosamine metabolites, N-nitrosobis (2-hydroxypropyl)amine (BHP) and N-nitroso(2-hydroxypropyl)(2-oxopropyl)amine (HPOP) [38,40]. In 1981, Pour [15] reported for the first time that application of BOP induced tumors in the prostate dorsal lobe in MRC rats. Some years later, the same author conducted a study to evaluate the effects of dietary fat on the development of prostatic cancer in Wistarderived MRC rats: testosterone propionate (100 mg/kg of body weight, subcutaneous injection) was daily administered for five days, and BOP (20 mg/kg body weight subcutaneous injection) was daily given to rats for three days, beginning with the third testosterone injection. This study lasted 72 weeks and showed that dietary fat did not influence the patterns of prostatic cancer [41]. Despite this, the BOP only induced squamous cell carcinomas development in the ventral prostate lobe, the only lobe that does not have a human homologue, making it an animal model not widely used [14].

N-methyl-N-nitrosourea (MNU)
The MNU is a nitro-compound that acts as direct carcinogen agent by methylation of the guanine nucleosides without previous metabolic activation [42,43] [46]. This method, although complex in execution, allows a higher incidence of rat prostate cancer. To achieve the maximal tumor incidence, the experimental work should be conducted until 50-60 weeks after MNU injection [9]. Over the years many authors used this method to study prostate carcinogenesis [47].
The animal models of chemically-induced prostate cancer makes possible the evaluation of the chemopreventive effect of different compounds. Chemoprevention may be defined as the use of natural, synthetic or biological agents to reverse, suppress or prevent initial phases of carcinogenesis or progression of malignant cells to invasive disease [32,48]. The studies may be consulted in Table 1.

3,2'-dimethyl-4-aminobiphenyl (DMAB)
The DMAB is a classical polycyclic aromatic hydrocarbon with carcinogenic properties for multiple organs, such as colon, urinary bladder, mammary glands and Zymbal glands [2]. This compound needs to be previously activated in the liver. Then, the metabolites may be also metabolized and their oxidation products interact with DNA causing transversions in nucleotides, inducing irreversible changes [49].  [53]. The DMAB method consists of a subcutaneous administration, normally at dose of 50 mg/kg/body weight, 10 times at 2-week intervals [52]. Silastic tubes filled with testosterone propionate may be used to promote tumor development. The implants should be placed subcutaneously and replaced at 6-week intervals [54]. In a general way, chemopreventive studies with this compound last 60 weeks, on average [55].
More details may be consulted in Table 2.

2-amino-1-methyl-6-phenylimidazol[4,5-b]pyridine (PhiP)
The PhiP is a heterocyclic amine isolated from cooked fish and meat. This compound may be metabolized to biologically active metabolites (N-hydroxy-PhIP and N-acetoxy-PhIP) that form DNA adducts [56][57][58] and induce cancer development in mammary gland, intestine and prostate [59,60]. Prostate cancer histopathological features are similar to those induced by DMAB. The rats exposed to PhiP develop prostate carcinomas in the ventral lobe, but not in the dorsolateral or anterior lobes. The experimental protocols using PhiP as prostate carcinogen consists of the administration of this compound mixed into the diet [32,59] or administrated intragastrically by gavage (70-200 mg/kg) [56,61,62]. This model can be used for chemoprevention studies. Detailed information concerning to these studies may be consulted in Table 3.

Hormonally-induced prostate cancer
Since the rat prostate is an androgen-sensitive tissue, the administration of testosterone may induce the development of invasive adenocarcinomas in dorso-lateral and in anterior prostate, with low incidence in various rat strains [9, 14,32]. The association of testosterone with estrogen, like 17β-estradiol, induces a higher carcinoma incidence in the dorsal, lateral and anterior prostate in Noble and Sprague-Dawley rats [14]. The methodology to induce prostate tumors with this compound consist of the use of silastic tubes filled with testosterone (crystalline testosterone or testosterone propionate) alone or in combination with estradiol (in two silastic tubes separated) implanted under the skin [14]. The wall thickness determines the speed of hormone release, as well as, the length of the tubing over which the hormone can diffuse out. The amount of the hormone in silastic tubes determine the duration of release. This model may be used in chemopreventive studies [63][64][65], but it is not very advantageous because of the early death of the rats due to the concomitant development of estrogen-induced pituitary tumors, which prevent the prostate carcinomas from growing and metastasis [14].

Implantation of cancer cell lines
Prostate cancer cells lines may be implanted in rats for tumor development. The cell lines may be either obtained from human prostate tumors and implanted into rats (xenograft model) [66], or obtained from chemically-induced or spontaneous rat prostate cancer (syngeneic model) [67][68][69][70][71]. Furthermore, these models may be orthotopic, if the prostate cancer cells are implanted in the tumor site of origin, in this case the prostate [72][73][74]; or heterotopic, if the cancer cells are implanted in a different place, for example subcutaneously [68] or in subcutis of the dorsal surface of the rat tail [71].
PAIII cancer cell line is an example of syngeneic model. This androgen insensitive cancer cell line is derived from a spontaneous prostate carcinoma from a Lobund-Wistar rat and it has the ability to form large primary tumors when injected subcutaneously [68,75]. Harvey Pollard and colleagues [68] used this model to test the influence of ascorbic acid on tumorigenesis and concluded that pharmacological doses of this compound may suppress tumor growth and metastasis. PLS 10 is another cancer cell line, it was established from DMAB plus testosterone-induced carcinomas in the dorsal prostate of male F344 rats, which may be implanted into rats [67]. The PSL10 cell line was used as a syngeneic heterotopic model into the flank of F344 rats in a study aiming to understand hyperthermic effects of magnetic particles on rat prostate cancer. The authors concluded that hyperthermia is an effective therapy for prostate cancer [67].
The xenograft prostate cancer models in rats are not widely used in experimental research due to the few immunodeficient rat strains available for use. Highlight the work of Tumati [66] that developed an advanced orthotopic prostate cancer model using a modified human prostate cancer cell line (PC3) implanted into the prostate of nude or Copenhagen rats to assess the effects of high-dose image-guided radiation therapy combined with biological agents [66]. In another work, Andressen [76] used immunodeficient male homozygous Sprague-Dawley rats injected with CRW22 human prostate cancer cell line to investigate the ability to induce bone metastasis.

Dunning model
Most of the in vivo rat tumor lines for testing new drugs for the treatment of hormonedependent prostate cancers originated from the Dunning R3327 adenocarcinoma of the Copenhagen rat. This cell line was developed from a spontaneous dorsal prostatic adenocarcinoma in a Copenhagen male rat in 1963. The tumor was identified in a necropsy and grafts of the tumor were transplanted into syngeneic rats and F 1 hybrids of a Copenhagen x Fischer cross. The transplanted tumor was histologically, biochemically and biologically similar to the rat dorso-lateral prostate [35]. Since then, several sublines were developed from Dunning R3327, such as AT-1, MAT-Ly-Lu, AT-4-R3327-5 and PAP [35]. Each subline has different characteristics and exhibits a range of phenotypes that mimic aspects of the man prostate cancer, like the slow-growing androgen-responsive tumor [7]. These cell lines may be passed in vivo without lose tumorigenicity and metastatic ability [35]. Since the tumor associated with this model is slow-growing, non-metastatic, androgen-responsive, and maintained the histologic appearance and biochemical properties of the rat dorsal prostate gland, it is considered an appropriate model for cancer research, namely for chemoprevention studies [35] ( Table 4).
The Dunning model has the main advantage of not requiring exposure to chemical compounds for its execution, thus reducing the health risk for the investigator, the environmental effects and costs.

Genetically-engineered models
Both chemically or hormonally-induced prostate cancer models are labor intensive and need a long period of latency for tumor development: at least 60 weeks for the DMAB and PhIP models, 50 weeks for the MNU models, 70 weeks for BOP models and 40 weeks for hormonally-induced models on average. As an attempt to overcome this disadvantage, genetically-engineered models were developed [77,78]. androgen-dependent [78], and develop mostly in the ventral, lateral and dorsal prostate lobes at high incidence [32,78]. The initial lesions in prostate epithelial cells may be observed at four weeks and the carcinomas become extensive at 15 weeks [32], providing a useful animal model to investigate the mechanisms underlying prostate cancer development, as well as, the effects of modifying factors. Another transgenic rat prostate cancer model was developed in the Lewis rat, crossing the Sprague-Dawley SV40-T rats with Lewis strain to study prostate cancer immunotherapy (Table 5). The prostate tumors developed with 100% penetrance and were androgen sensitive [79].

Follow-up of animal models -prostate imaging
Although the prostate cancer remains the fifth cause of death by cancer among men, the mortality associated with this type cancer has decreased in the past decades, mainly due to the widespread use of screening strategies. Despite the only definitive way to confirm prostate cancer is through a prostate biopsy (frequently ultrasound guided biopsy) [80], screening for this kind of cancer includes digital rectal examination (DRE) focused on prostate size and consistency, or more typically, a change on serum prostate-specific antigen (PSA) [81]. Furthermore, the prostate changes may be non-invasively evaluated through imaging modalities, such as ultrasonography (transrectal ultrasound) [82]. Nowadays, ultrasonography provides not only anatomic information (prostate dimensions and macroscopic structure of parenchyma) through the use of Bmode (grayscale ultrasonography) [66,83], but also functional information on tumor microenvironment (evaluation of prostate vascularization) through the use of Power Doppler, Color Doppler, Pulsed Doppler, B Flow, and ultimately contrast agents.
Contrast-enhanced ultrasound (CEUS) is based on the intravenous administration of a contrast agent [80] and allows not only to quantify and evaluate the pattern of distribution of the vessels inside the parenchymal prostate (tortuous vessels, centripetal/centrifugal enhancement of contrast agent), but also to perform dynamic studies, evaluating the arrival time of the contrast agent, peak intensity, time to peak, ascending-slope, descending-slope and area under the curve [84].
When compared with other imaging modalities, the ultrasonography has some advantages, namely the ultrasound apparatus are portable, they are less expensive, they are recommended for claustrophobic patients and, the most important one, does not impose ionizing radiation. Although the ultrasonography is the oldest and the most widely used imaging modality for prostate imaging [80], it is limited on the detection and location of prostate cancer, and on the tissue contrast between benign and cancerous tissues due to its lower sensitivity and specificity [82]. To overcome these limitations, other imaging modalities, such as Magnetic Resonance Imaging and Positron Emission Tomography have been employed on prostate screening.
Magnetic Resonance Imaging allows the visualization of the prostate contours and internal anatomy, providing functional and structural information of tumor vasculature and physiology [84] with great sensitivity, but with low specificity [85]. This imaging technique is usually combined with other techniques, like Diffusion Weighted Imaging that improves the detection and localization of prostate cancer; Magnetic Resonance Spectroscopy that assesses the relative concentration of different chemical compounds in tissue, and Dynamic Contrast-Enhanced Magnetic Resonance imaging whose principles are based on tumor angiogenesis [85,86]. Positron Emission Tomography is a non-invasive imaging modality that uses radiolabel tracers and gamma cameras to measure sensitively and quantitatively the concentration of these radioactive molecules in the prostate [84]. This imaging modality is often combined with Magnetic Resonance Imaging for anatomical localization of the spots for radiotracers [80]. Positron Emission Tomography is more used in later stage cancer than in diagnosis, being useful on the detection of biochemical relapse, cancer recurrence [80] and metastasis. This imaging modality has the disadvantages of not detect small lesions, not distinguish benign and malignant processes, the higher costs, and requirement of advanced training and skills [87]. It is worth to note that these imaging modalities may also be used in rat models for prostate cancer diagnosis and to monitor the effects of new potential therapies for cancer treatment [88], however the published works are scarce.
Our research team has some ultrasonographic, Magnetic Resonance Imaging and administration (more evident on ventral lobes) ( Fig. 2A and B). An inverse effect was observed six weeks after the MNU administration, with the carcinogen promoting the increase of prostate volume (with higher incidence on ventral lobes) and compressing the neck of urinary bladder ( Fig. 2C and D). The prostate volume continued increasing overtime, until the end of the experimental work, with a marked volume enhancement of dorsal prostate lobes and the vesicular glands placed near the prostate lobes ( Fig. 2E and F). We also performed a study of prostate vascularization using Power Doppler, B Flow and contrast-enhanced ultrasound. An increase on prostate vascularization was observed between the 35 weeks and 61 weeks after MNU administration. Some Magnetic Resonance Imaging and Computed Tomography studies were also performed in normal and cancerous prostates (Fig. 3).

Sample collection and histological evaluation
A standardized protocol to collect prostate tumors was not yet established. Some researchers remove accessory sex glands together with urinary bladder and separate the different prostate lobes from urinary bladder after fixation in formalin [19,63,89].
Other researchers remove the accessory sex glands, fix them in formalin and cut them into slices (where include urethra and seminal vesicles) to paraffin inclusion [27,31,62,90,91]. Bosland suggests that the accessory sex glands are best removed and fixed together with urinary bladder and then they should be separated [9]. According to our experience, this seems to be the most appropriate method to identify the different prostate lobes.

Spectrum of prostate lesions
As mentioned elsewhere, cancer is a multifactorial disease and factors that are responsible for prostate tumorigenesis remain largely unknown. As prostate gland is an endocrine-responsive tissue, many studies focused on the effect of androgens, estrogens and their metabolites on prostate tissues.
Over the years, several rat prostate models using chemical carcinogens have been established. In 1977, Fingerhut and Veenema [92] reported carcinomas induced by DMBA in gonadectomized animals. Later in 1983, Pour described hyperplastic and metaplastic lesions, and squamous cell carcinomas, on MCR rat ventral prostate after repeated BOP administration [93]. These lesions in the ventral prostate, which has no homologue in humans [36,94,95], occur spontaneously in some rat strains, especially in old animals, while chemical carcinogens induce neoplastic changes predominately in the dorsolateral lobe [96,97] .
Carcinomas in situ of the dorsolateral prostate (an embryological homologous to the human prostate), were reported in other chemical induction models using DMABP [16] or a combination of DMABP and ethinyl estradiol [98]. Carcinomas of the dorsolateral prostate were also described by  on the Lobund-Wistar rat, after a single MNU administration [99]. Bosland and Prinsen (1990) [95] using MNU or DMBA, following chemical castration and testosterone propionate pretreatment, induced invasive adenocarcinomas and carcinomas in situ of the dorsolateral prostate on Wistar rats, 63 weeks after carcinogen injection, mainly in the MNU treated animals. They also reported reactive hyperplasia of the dorsolateral prostate associated with acute and chronic inflammatory processes as a common finding in all experimental group. Apparently, reactive hyperplasia is a consequence of inflammation [100,101].
Later in 1998, McCormick and colleagues [102] reported a high incidence of accessory sex glands cancer, mostly in dorsolateral and anterior prostate lobes, in rats treated sequentially with MNU and testosterone, but lower cancer incidence in rats receiving only MNU or treated with testosterone only; no lesions were observed in untreated animals, suggesting that carcinogenesis depends on both chemical and hormonal stimulation.
In 1999, Rao et al. [19] also described dorsolateral and anterior prostate carcinomas in Wistar-Unilever rats using cyproterone acetate and testosterone propionate, followed by a single dose of MNU in animals with chronic androgen stimulation. They also reported that some animals developed seminal vesicle tumors. Previously, Lucia et al.
(1995) [91] reported a high incidence of seminal vesicle tumors, rather than prostate tumors, induced by MNU and testosterone propionate in Lobund-Wistar rats.
More recently, in 2014, Bosland [46] reported a high frequency of prostate carcinomas after MNU injection in rats with subcutaneous testosterone implants. The frequency was lower in rats exposed to testosterone alone (not exposed to MNU) and tumors did not occur in rats given MNU or not treated. Based on these findings, Bosland [46] suggested that testosterone is a carcinogen and a tumor-promoting agent.
Our group studied the influence of chemical carcinogen and hormonal stimulation on the induction of prostate cancer on Wistar rats using a multistep experimental protocol where the animals were submitted to a chemical castration through the administration of flutamide, followed by the administration of testosterone propionate, a single injection of the carcinogen MNU and a chronic exposure to testosterone using subcutaneous implants. Twenty-three weeks after the beginning of the experiment, we observed mainly prostate hyperplastic lesions, and occasionally low grade dysplastic lesions, mostly in dorsolateral prostate, which also showed acute inflammation of the acini, focal necrosis and reactive hyperplasia, with small focal areas of chronic stromal inflammation. In fact, all groups of the experiment showed acute and/or chronic inflammation. Inflammation was also frequently reported by other studies [96,101] and reported as more common and severe in the dorsolateral prostate [101], as we observed. Some animals also developed atypical hyperplasia of anterior prostate and seminal vesicle (Fig. 4).
Rats sacrificed 49 weeks after the beginning of the study showed more commonly dysplasia, in situ carcinoma or invasive carcinomas. Our data suggest that chronic exposure to testosterone acts as a tumor promoter in cells initiated by MNU, fostering the transition from hyperplasia to dysplasia and finally the progression from in situ to invasive carcinoma, being this promoter effect apparently time-dependent.

Conclusions
Experimental data concerning to the rat models of prostate carcinogenesis was reviewed in this work. Although several animal models are available to study prostate cancer and a perfect model does not exist, they provide an important tool to study human and animal prostate carcinogenesis, and to evaluate the effects of potential preventive and therapeutic strategies. The model should be chosen by the researchers, taking into account the aims of their studies, the costs, and the advantages and disadvantages of each one.
Although complex, time-consuming and labor-intensive, the model of prostate cancer hormone and chemically-induced in male rats is the most frequently used due to its advantages when compared with the remaining models. When applying the sequential treatment with an anti-androgen, the carcinogenic agent MNU and the chronic treatment with silastic implants filled with testosterone, a high incidence of prostate carcinomas is observed in several rat strains, with the maximal incidence reached 50-        Death of tumor cells [129] PNU 157706 ( (N-(1,1,1,3,3,3hexafluorophenylpropyl)-3-oxo-4aza-5a-androst-1-ene-17bcarboxamide)) and flutamide (5α-reductase inhibitor ;nonsteroidal antiandrogen Oral (10 mg/kg/day of PNU 157706 and 1 or 5 ml/kg/day of flutamide 6 days/week) for 9 weeks Inhibited prostate cancer of both compounds [130]  Laser interstitial thermotherapy (LITT) LITT (980nm diode laser for 75s) after 3 wks of tumor inoculation Induced tumor necrosis [134]