Transplantable Animal Studies and Whole-Body Optical Imaging in Prostate Carcinoma

Abstract

Current treatments of advanced prostate cancer only marginally increase overall survival and can be regarded as predominantly palliative. Hence, there is an urgent need for novel therapeutic strategies for the treatment of primary tumors and, more importantly perhaps, for the prevention of tumor progression and metastasis formation. Clinically relevant preclinical models are therefore urgently needed. An ideal, clinically relevant preclinical model would mimic the genetic and phenotypic changes that occur at the different stages of human prostate cancer progression and subsequent metastasis. In this chapter, transplantable xenograft prostate cancer models are described, in which human prostate cancer cells are transplanted into host animals (e.g., immune-deficient mice). Cancer cells can be administered to the small laboratory animals in various ways, including inoculation of the prostate tumor cells subcutaneously, at the anatomical site of origin (orthotopically), or at the metastatic site. In addition, we describe imaging methods suitable for small laboratory animals with emphasis on optical imaging (bioluminescence and fluorescence).

1 Introduction

1.1 Clinical problem of Prostate Carcinoma

Prostate carcinoma is the most common cancer in males and the second leading cause of cancer death in the Western world. Due to an expanding population at risk, the socioeconomic and medical impact of prostate carcinoma is increasing. Current treatments of primary prostate tumors are initially very effective. However, in 20–30% of newly diagnosed patients with apparent organ-confined prostate cancer, beneficial responses are followed by tumor recurrence at distant sites leading to incurable, devastating metastatic disease.

Current treatments of advanced prostate cancer only marginally increase overall survival and can be regarded as predominantly palliative [1, 2]. Hence, there is an urgent need for novel therapeutic strategies for the treatment of primary tumors and, more importantly perhaps, for the prevention of tumor progression and metastasis formation. Clinically relevant preclinical models are therefore urgently needed.

1.2 Preclinical Models

To establish clinically relevant models to predict therapy response and metastasis formation, these models ought to mimic both the genetic and epigenetic changes that occur at the various stages of human prostate cancer progression and subsequent metastasis. Unfortunately, no such models yet exist for small laboratory animals, due to the lack of spontaneous metastasis, the very long latency of metastasis, or the presence of only intravascular metastases [3, 4]. As a result, numerous preclinical animal studies have been developed, in which prostate cancer cells are transplanted into a host animal (i.e., transplantable tumor models). In syngeneic models, the inoculated cells are of the same species and genetic background [5], whereas in xenograft models, the cancer cells are of human origin inoculated into immune-deficient mice, including BALB/c nu/nu nude and severe combined immune-deficient (SCID) mice. A disadvantage of transplantable animal models is that only specific stages of the metastatic cascade are characterized (e.g., no spontaneous tumor formation), and some crucial features of the interaction between tumor cells and the surrounding tumor microenvironment might be altered in these models. In addition, the syngeneic inbred mouse models lack genetic complexity, whereas the xenograft preclinical models have an incomplete immune system.

1.3 Transplantable Animal Models

In the transplantable animal models, various methods exist to inoculate the prostate cancer cells, including subcutaneous inoculation, orthotopic inoculation (i.e., at the anatomical site of origin—the prostate), or inoculation at the metastatic site (the bone/bone microenvironment). Moreover, cells can be administered via the left cardiac ventricle to mimic metastatic spread of the tumor cells throughout the body (systemic inoculation).

1.4 Subcutaneous Transplantation

Subcutaneous administration is a straightforward method of inoculation and remains a valuable approach for high-throughput drug screening. However, to more accurately studying the processes of tumor progression and metastasis, it is important to administer prostate cancer cells to a more biologically relevant environment such as the prostate or to the metastatic site.

1.5 Orthotopic Transplantation

Administration of prostate cancer cells into the prostate, the tissue from which the tumor cells were originally derived, is called orthotopic inoculation. Orthotopic inoculation of an established prostate cancer cell line, PC-3M-Pro4 results in tumor growth in the prostate, as well as metastasis towards the locoregional lymph nodes. However, no reliable bone metastases have been formed in this model [6,7,8]. Current drawbacks of orthotopic transplantation include the lack of distant metastasis formation and the possibility of tumor cells leaking into the peritoneum following surgery as well as the trauma of opening the mouse peritoneum itself.

1.6 Intraosseous Transplantation

Another transplantable animal model consists of inoculation of prostate tumor cells into one of the metastatic sites, the bone/bone marrow microenvironment (intraosseous or intra-bone inoculation). This technique can be used to study the interaction between the cancer cells and the metastatic environment (stromal compartment), recapitulating the interactions in established prostate cancer bone metastasis. Intraosseous inoculation results in either osteolytic or osteoblastic lesions or a mixture of those, depending on the prostate cancer cell line used. For example, osteoblastic lesions are formed after inoculation of C4-2B, MDA-PCa-2b, LAPC-9, and LuCaP 23.1 prostate cancer cell lines [9], whereas the human prostate cancer cell lines PC-3, Du-145, and RM-1 cells predominantly cause osteolytic lesions [7]. Key factors in the process of tumor-induced changes in the bone such as MMP-7 and MMP-13 have been identified using inoculation of mice with these established cell lines [10, 11]. Drawbacks of these models are the initial bone damage and resulting repair mechanisms due to the procedure as well as potential inflammation at the injection site. In addition, since the tumor cells have not entered the bone microenvironment via the endothelium (one of the early steps of the metastatic cascade), the precise location of the tumor cells might be different (i.e., the tumor cells potentially do not reside at their specific niche) [12, 13].

1.7 Systemic Inoculation of Cancer Cells

To study the process of homing to distant target organs and subsequent growth at the metastatic site, systemic inoculation of tumor cells either in the lateral tail vasculature or in the left cardiac ventricle are widely used experimental metastasis models [6, 14,15,16,17,18]. Using this technique, the later stages of prostate cancer metastasis can be examined, mainly the ability of the cancer cells to home to the metastatic site, survive in the stroma of the metastatic target organ, proliferate and finally form an overt metastasis. Moreover, these models are valuable in screening the response to compounds targeting formation and/or growth of prostate cancer metastases. Inoculation in the tail vasculature predominantly results in pulmonary metastasis, whereas bone metastasis is more prevalent when tumor cells are inoculated in the left cardiac ventricle. A potential disadvantage of these models is that early steps in the metastatic cascade are bypassed. Advantages of these models include control on both the amount of and characteristics of the cells that are inoculated. Using intracardiac inoculation, the tumorigenic and metastatic potential of different subpopulations of cancer cells can be investigated, such as cells with a stem/progenitor phenotype [19] or with altered expression levels of a gene of interest [20, 21].

1.8 Imaging of Small Laboratory Animals

Although it is essential to improve the preclinical models to more closely mimic the course of human prostate cancer, accurate imaging and data acquisition of the tumor progression and metastasis as well as the biological processes involved is also crucial for the strength of the experimental approach. To study molecules, pathogenic processes, drug delivery and response, several novel imaging technologies have been developed allowing real-time imaging in a preclinical and occasionally clinical setting [6, 22, 23]. Several imaging technologies have been developed for small animals, such as microcomputed tomography (μCT), micropositron emission tomography (μPET), single photon emission computed tomography (SPECT), magnetic resonance imaging (MRI), ultrasound imaging, and optical imaging (reviewed in [3, 4]). It is important to note that some imaging techniques are better suited for certain applications than others. For example, to monitor tumor cell biology, tumor burden, progression, and metastasis, highly sensitive approaches such as PET/SPECT and optical imaging (photo-acoustic imaging, bioluminescence and fluorescence imaging ) are more appropriate, while to obtain anatomical detail CT and MRI are more suitable [24].

Combining several imaging techniques, called multimodality imaging, may provide a better solution to overcome the limitations of the independent techniques. Multimodality imaging will expand as well as improve the available information for the preclinical models. In this chapter, we will focus on noninvasive whole-body optical imaging (encompassing bioluminescence and fluorescence), since optical imaging is sensitive, permits longitudinal and quantitative real-time gene expression, cellular localization, and drug response studies in small laboratory animals [3, 4]. In addition, optical imaging is the least costly, requires shortest imaging times of 5 min or less, is easy to use, and can image several animals at once.

1.9 Noninvasive Whole-Body Optical Imaging

Optical imaging is based on emission of light from labeled cells or probes (the source of the light being either fluorescent or bioluminescent ). Noninvasive, whole-body optical imaging enables examination of longitudinal and quantitative real-time gene expression, cellular localization, as well as drug response studies in small laboratory animals .

1.10 Bioluminescence

Bioluminescence imaging detects photons emitted by an enzymatic reaction in which luciferase catalyzes the conversion of d-luciferin into oxyluciferin in an ATP-dependent process. Bioluminescence imaging is a widely used imaging technique because of the low background, high signal-to-noise ratio, and the short acquisition time. Several luciferases have been detected, of which firefly luciferase derived from the firefly Photinus pyralis is the most widespread used (see Note 4 ). It has to be noted that bioluminescence imaging has limited depth information and limited spatial resolution.

1.11 Fluorescence

Using fluorescence imaging , reporter proteins (either fluorescent proteins or fluorochromes targeted to specific cell compartments or molecules) need to be excited by an external excitation source. Subsequently, fluorescence imaging detects the emission of light at specific wavelength from the reporter proteins. Fluorescence imaging can be divided into Reflectance Imaging (FRI) and Molecular Tomography (FMT), which can provide 3D information [25]. Like bioluminescence imaging, fluorescence imaging is restricted due to the limited penetration depth and spatial resolution. Moreover, autofluorescence of mammalian tissues is the most important source of background impairing the sensitivity of the fluorescence signal (see Note 7 ).

2 Materials

2.1 Cell Maintenance

For PC-3 cells and their derivatives (e.g., PC-3M-Pro4), Dulbecco’s modified eagle’s medium DMEM with 4.5 g/L glucose, GlutaMAX™ Supplement and pyruvate supplemented with 10% FCII (Fetal Clone II, Hyclone) and penicillin/streptomycin and 0.8 mg/mL geneticin (G418, neomycin) is used (see Note 1 ). For cell maintenance, cells are passaged at 80–90% confluence.

2.2 Constructs

For generation of stable cell lines, we used a modified pGL4 reporter construct (Promega) containing the mammalian optimized firefly luciferase 2 under control of the CAGGS promoter (CMV early enhancer/chicken beta actin CAG) with the neomycin selection marker (geneticin, G418) (see Note 2 ).

For the generation of stable reporter cell lines, we generated modified pGL4 reporter constructs with destabilized FFLuc2 under control of the appropriate promoter sequence (e.g., BRE₄ or CAGA₁₂ elements for respectively the BMP and TGFβ pathway).

2.3 Imaging Systems

For bioluminescence and fluorescence imaging of small laboratory animals , the Xenogen IVIS Lumina Series III Pre-clinical In Vivo Imaging System (Perkin Elmer) was used. The sensitive range of the CCD camera sets the wavelength range of the IVIS Lumina for fluorescence applications from 400 to 950 nm. For fluorescence imaging, CRi Maestro in vivo multispectral imaging system for fluorescence imaging can be used (see Note 3 ). The wavelength range of the Maestro for fluorescence applications ranges from 500 to 900 nm. For anesthetization of the mice, the Isoflurane Tec-3 Anesthesia Vaporizer System is utilized.

For analyzing the obtained results, Living Image software and Maestro software are used for respectively the IVIS Lumina and the Cri Maestro.

2.4 Luciferase Substrates

The substrate for Photinus pyralis (firefly) luciferase is luciferin (also known as d-luciferin, (d-(−)-(2-(6′-hydroxy-2′-benzothiazolyl) thiazone-4-carboxylic acid)), a monopotassium salt (Promega). The substrate for Renilla reniformis (Renilla) luciferase and Gaussia princeps (Gaussia) luciferase is coelenterazine (ViviRen™ Live Cell Substrate (Promega)). The substrate for NanoLuc is furimazine (Promega) (see Note 4 ).

2.5 Materials for Intraosseous Inoculation

• 1. Isoflurane Tec-3 Anesthesia Vaporizer System.

• 2. 70% EtOH.

• 3. Pair of bended tweezers.

• 4. Pair of tweezers.

• 5. Needles (25GA 5/8 0.5 × 16).

• 6. Insulin syringes.

• 7. PBS.

• 8. 1 mL syringes.

• 9. Surgical scissors.

• 10. Tissue paper.

• 11. Gauze pads.

• 12. Stainless steel wound clips 9 mm.

2.6 Materials for Orthotopic Inoculation

• 1. Isoflurane Tec-3 Anesthesia Vaporizer System.

• 2. 70% EtOH.

• 3. Pair of tweezers.

• 4. Needles (30G 1/2, BDMicro-Fine).

• 5. 1 mL syringes.

• 6. Surgical scissors.

• 7. Gauze pads.

• 8. Stainless steel wound clips 9 mm.

• 9. Growth Factor Reduced Matrigel (BD Biosciences).

2.7 Materials for Anesthesia

• 1. Isoflurane Tec-3 Anesthesia Vaporizer System.

• 2. 70% EtOH.

• 3. 1 mL syringes.

• 4. Needles (30G 1/2, BDMicro-Fine).

• 5. Gauze pads.

3 Methods

3.1 Cell Lines (See Note 5 )

For optical imaging using bioluminescence , tumor cell lines stably expressing luciferase can be used (see Notes 4 – 6 ). For optical imaging using fluorescence, preferably a red-shifted fluorochrome (see Note 7 ) is used. Fluorescence imaging can be performed with either stable expressing cell lines or labeling of the cells with a fluorochrome depending on the experimental questions (see Note 8 ).

Many bone metastatic subclones of established prostate cancer cell line exist that have variable propensity for bone colonization and growth. For our experiments we use a subclone of PC-3 cells, PC-3M-Pro4 that readily forms bone metastasis after inoculation in vivo.

Before inoculation, the tumor cells are maintained in their appropriate medium containing G418 (or appropriate antibiotic selection for the luciferase or fluorochrome) and passaged approximately every 72 h (1:10 dilution). Ideally, cells should be maintained in culture only briefly before inoculation (preferably less than 2 weeks). One day before inoculation, refresh the medium with medium without antibiotic selection. Tumor cells were harvested at 70–80% confluence with 0.15% trypsin/EDTA solution at 37 °C (keep the trypsinization time <5 min). Immediately inactivate the trypsin/EDTA with 10 mL ice-cold DMEM containing 10% FCII. Centrifuge the cells at 800 rpm for 7 min and resuspend in ice-cold PBS. Count the cells and immediately store the cells as a single-cell suspension in PBS on ice before inoculation in vivo. It is important to have a single-cell suspension, since clumps of cells can result in an embolus in the mice. Moreover, time between trypsinization and the inoculation of the cells should not exceed 30 min.

3.2 Experimental Procedures on Animal Models

3.2.1 Mice (See Note 9 )

Female/male athymic nude mice (Balb/c nu/nu) were purchased from Charles River (L’Arbresle, France) and housed in individual ventilated cages under sterile conditions; sterile food and water were provided ad libitum. Animal experiments were approved by the local committee for animal health, ethics, and research, and carried out in accordance with European Communities Council Directive 86/609/EEC. For the different procedures, mice of different age were utilized as indicated.

3.2.2 Anesthesia

Mice were anesthetized before surgical and analytical procedures were performed. In the intracardiac inoculation model, mice were anesthetized with an induction dose of 3% isoflurane and 2–3 L/min O₂ and maintained under anesthesia at a dose of 1% isoflurane and 2–3 L/min O₂. In the intraosseous inoculation model and the inoculation of the cells at the orthotopic site, mice were anesthetized with an induction dose of 3% isoflurane and 2–3 L/min O₂ and maintained under anesthesia at a dose of 2–3% isoflurane and 2–3 L/min O₂.

3.2.3 Intraosseous Inoculation

• 1. Anesthetize the mouse (7-week-old) with isoflurane.

• 2. Apply 70% EtOH solution to a gauze pad and disinfect the legs.

• 3. Lift the skin and cut an incision of approximately 5 mm, at the broad part of the tibia.

• 4. Clean the tibia by removing the corium. This can be done by scraping the tibia with a 25GA 5/8 0.5 × 16 needle.

• 5. Fixate the tibia with a bended tweezers in one hand. Keep your hand and the tweezers on the table for stable fixation.

• 6. Drill a hole with the 25GA 5/8 0.5 × 16 needle at the start of the broad part of the tibia by rolling the needle between your fingers using a steady and firm pressure.

• 7. Drill the second hole, with a new needle, directly under the white part of the bone, just under the knee.

• 8. Flush the bone marrow by injecting PBS in the upper hole, pointing the syringe in the flow direction. The bone marrow will flow out of the lower hole, slightly pink and cloudy.

• 9. Inject 10 μL of the tumor cells (a single-cell suspension of 10⁵ cells/10 μL of PBS) with an insulin needle in the upper hole, pointing the syringe in flow direction.

• 10. Remove overload of tumor solution with a gauze pad, but avoid the drilling holes.

• 11. Close the wound with a stainless steel autoclips (Leica Biosystems, Nussloch, Germany), the clip can be removed under anesthesia, after 3–5 days.

3.2.4 Intraprostatic Inoculation

• 1. Anesthetize the mouse (6–7-week-old) with isoflurane.

• 2. Apply 70% EtOH solution to a gauze pad and disinfect the abdomen.

• 3. Lift the skin and cut an incision.

• 4. Put aside the bladder, and in case of fat depositions on the bladder and prostate, remove the fat depositions (see Note 10 ).

• 5. Inject 10 μL of the tumor cells (a single-cell suspension of 10⁵ cells/5 μL of PBS combined with 5 μL growth factor reduced Matrigel (BD Biosciences)) with an insulin needle into the dorsal prostate pointing the syringe in the direction of the lobe.

• 6. Close the wound with a stainless steel autoclips (Leica Biosystems, Nussloch, Germany), the clip can be removed under anesthesia, after 3–5 days.

3.2.5 Intracardiac Inoculation

• 1. Measure weight of the animals before intracardiac inoculation. Mice need to be at least 11 g for sufficient survival of the mice.

• 2. Anesthetize the mouse (4–5-week-old) with isoflurane (see Note 11 ).

• 3. Place the mouse on its back, facing the mouse lateral.

• 4. Apply 70% EtOH solution to a gauze pad and disinfect the chest.

• 5. Place your index finger and thumb under the ribs of the mouse and let the mouse rest on them.

• 6. Slightly lift the mouse.

• 7. Insert the needle in the cavity between the xiphoid cartilage and the ribs, keeping the syringe even with the body.

• 8. Inject directly through the skin, into the left ventricle of the heart.

• 9. For assessment of appropriate inoculation into the left cardiac ventricle, check whether blood enters the syringes in pulses.

• 10. Slowly inject a single-cell suspension of 10⁵ cells/100 μL in PBS into the left cardiac ventricle using a 30-gauge needle (30G 1/2, BDMicro-Fine; Becton Dickinson) (see Note 12 ). Injecting too fast can kill the animal instantly due to the arteries being clogged. Be careful to avoid significant movement of the needle to prevent puncturing the heart or spilling cells into the chest cavity.

• 11. After injecting 100 μL, keep the needle in place for 5 s before removing it.

• 12. Anesthetized animals were placed in the IVIS100 Imaging System (Xenogen, Alameda, CA, USA) within 60 min after injection and imaged dorsally and ventrally 5 min after intraperitoneal injection of d-luciferin. A successful intracardial injection was indicated on day 0 by images showing systemic bioluminescence distributed throughout the body of the mouse (Fig. 1b). A technical failure will be noticeable by bioluminescence signal in the heart area only (Fig. 1a). Only animals with evidence of a satisfactory injection continued the experiment.

• 13. Assessment of subsequent metastasis was monitored in vivo once or twice a week, depending on the tumor model, for up to 42 days.

Fig. 1

Bioluminescent imaging of PC-3M-Pro4luc2 cells after intracardiac inoculation. PC-3M-Pro4luc2 cells were inoculated in the left cardiac ventricle and after several time points after inoculation bioluminescence was measured with IVIS Lumina (Caliper Life Sciences). (a) Representative image of a mouse with pericardial inoculation at 1 h, 3 h, 24 h, and 20 days after inoculation. (b) Representative images of mice with intracardiac inoculation at 1 h, 3 h, 24 h, and 20 days after inoculation

3.2.6 Bioluminescence Measurements

Outgrowth of spread of the tumor cells was monitored (number of measurements depending on the experiment) by whole-body bioluminescent imaging (BLI) using an intensified charge-coupled device (CCD) video camera in a light-tight specimen box of the in vivo Imaging System (IVIS100/Lumina; Xenogen).

Animals were anesthetized with isoflurane and were given d-luciferin (15 mg/mL in D-PBS; 25 μL/10 g body weight) by intraperitoneal injection. Mice were placed into the light-tight camera box with continuous exposure to isoflurane. Imaging time ranged from 10 s to 1 min (generally 30 s), depending on the tumor model, 5 min after d-luciferin injection (see Note 13 ). Three mice were imaged each time from a dorsal and a ventral view. The photons emitted from the bioluminescent tumors or cells were detected by the IVIS camera system, integrated, digitized, and displayed. When saturated pictures are taken, first the F-stop is adjusted (diaphragm). If the image is still saturated, the image time can be adjusted (see Notes 14 and 15 ).

3.2.7 Evaluation of Bioluminescence Using Living Image

Quantification of signals was performed by the acquisition and analysis software Living Image® (Xenogen). Values are expressed as RLUs in photons/second. Numbers of tumors per animal were counted by eye from a dorsal and ventral view with particular regard to the singular count of one metastatic lesion.

3.3 After the Experimental Period

The human endpoints for the animal experiments must be taken into account (human endpoint is the point at which an experimental animal’s pain and/or distress is terminated, minimized, or reduced by taking actions such as euthanizing the animal, terminating a painful procedure, or giving treatment to relieve pain and/or distress). In our experiments, the human endpoints are: (1) Inability to access food or water. (2) Rapid or progressive weight loss (in 1–2 days more than 15% weight loss, or more than 20% vs. the weight of the mouse at the beginning of the experiment). (3) Severe circulatory or respiratory problems. The following symptoms may arise: increased breathing or heart rate, difficulty breathing, cold and blue extremities or ears, systemic edema. (4) The behavior and movements of the animals are significantly different. The following symptoms may rise: passive or hyperactive behavior, self-injury (self-mutilation), abnormal muscle tone including cramps. (5) Dehydration of the animals, the skin loses its elasticity. Skin pinched over the back should return to its normal position after it is released. In a dehydrated animal, the skin will remain tented. (6) Ulcerated, necrotic, or infected tumors. (7) Tumor burden greater than 10% body weight (>10% of normal body weight or 2 cm³). (8) Tumors that interfere with eating or impair ambulation.

In case of human endpoints or the beforehand set experimental endpoints, mice were anesthetized and sacrificed by cervical dislocation. Tumors were dissected and processed for either further histological analysis or RNA isolation for gene expression profiling.

Evaluation of the metastasis should be performed by immunohistochemistry (see Note 16 ): For example, bone metastases in a subset of mice were also examined by Goldner staining after mice were sacrificed using decalcified bone.

4 Notes

• 1. Each cell line has its own special medium supplemented with various antibiotics and serum, also depending on the laboratory. In our laboratory, FCII is used since it is a fetal bovine serum (FBS) alternative, containing additional growth factors and supplements. The benefit of using FCII is the lower variability in batches compared to different fetal bovine serum batches. PC-3M-Pro4 cells are PC-3M cells passaged four times in the prostate of Balb/c nu/nu mice to generate highly metastatic clones that are capable of readily forming bone metastases.

• 2. Several constructs are available for bioluminescence measurements, of which some are optimized for use in mammalian cells to reduce background signals such as the pGl4 vectors (Promega). In addition, the intensity of the luciferase signal critically depends on the promoter in front of the luciferase gene. For example, the CMV promoter tends to be hypermethylated in vivo [26]. In addition, some promoters react to certain signaling pathways or therapeutic compounds, in that case it should be taken into consideration that the luciferase signal could be different compared to the tumor burden. For in vivo experiments with established prostate cancer cell lines, we prefer to use EF1α or CAGGS promoters, since these promoters appear to be stable in vivo and not responsive to major signaling pathways expressed in the prostate and bone microenvironment (such as the TGFβ superfamily).

• 3. Other possible equipment include the IVIS100 Imaging System (Perkin Elmer) capable of bioluminescence imaging and the IVIS® Spectrum in vivo imaging systems (Perkin Elmer). The latter can, like the IVIS Lumina, be used for multimodal imaging (both fluorescence and bioluminescence). The IVIS spectrum is, depending on the spectrum version, suited for multimodal 3D tomography combined with integrated X-ray and μCT. This provides the possibility to project the optical data sets on data acquired with other imaging techniques such as MRI or μCT.

  However, for fluorescence imaging , CRi Maestro multispectral imaging system is preferred due to the inbuilt multispectral imaging software algorithms which enable rapid imaging and quantification of more than one fluorescent signal in real time. The employed spectral unmixing enables imaging of multiple fluorescent signals simultaneously together with the removal of the majority of the autofluorescence from the image (depending on the fluorescent proteins used, see also Note 7 ). The increase in sensitivity by using multispectral imaging is approximately 300-fold, and separating the autofluorescence from the label fluorescence improves quantitation and lowers the detection limit noticeably. All machines include heated platforms to keep the mice at normal body temperature since there will be less light emission if anesthesia has decreased the animal’s body temperature.

• 4. Bioluminescent imaging is compromised by low signal intensity, impairing imaging of cells in deeper tissues and thereby restricting the sensitivity of the imaging method. Choosing the optimal luciferase for the designed experiments is therefore essential. For cell tracking, stable and high expression of the luciferase is essential. Firefly luciferase, derived from the firefly Photinus pyralis, which catalyzes the substrate luciferin, is the most extensively used luciferase enzyme in cell-based bioluminescent imaging [27]. The luminescent reaction is ATP-dependent. A major improvement has been made with firefly luciferase 2, a mammalian codon-optimized firefly (see Note 5 ). NanoLuc® (Nluc) luciferase is a small enzyme (19.1 kDa vs. 61 kDa-sized firefly luciferase) about 100-fold brighter than either firefly or Renilla reniformis luciferase using a novel substrate, furimazine, to produce high intensity luminescence [28]. The luminescent reaction is ATP-independent and designed to suppress background luminescence for maximal assay sensitivity (Invitrogen). Still, the detection of NanoLuc in deep tissues is limited due to the emission of predominantly blue light by this enzyme. Renilla luciferase, derived from the anthozoan sea pansy Renilla reniformis, catalyzes coelenterazine, a substrate distinct from luciferin. However, the absorption properties of tissues limit its use in vivo [27]. The humanized Gaussia luciferase enzyme, derived from the copepod Gaussia princeps, likewise uses coelenterazine as a substrate and does not require ATP, but emits a markedly more intense signal compared to Renilla luciferase and may therefore overcome the limitations associated with Renilla luciferase. It has to be taken into account that the native Gaussia luciferase enzyme is secreted, markedly attenuating the in vivo bioluminescent signal of cells that express the enzyme. Therefore, Gaussia luciferase can be used as a biomarker for monitoring tumor progression and treatment response of systemic metastases rather than for cell tracking. A blood Gaussia luciferase assay revealed early growth of metastatic tumors before BLI could visualize their presence [29]. Several versions of the Gaussia luciferase exist, including a transmembrane version, that can be used for cell tracking purposes. Other, naturally occurring, variants of luciferase (Chromo-Luc™, Promega) were obtained from click beetles (Pyrophorus plagiophthalamus), and produce light at several different wavelengths of emission (from blue to red). Based on the proposed studies, one or several luciferases can be chosen for cell tracking experiments. For reporter studies , e.g., examining expression levels of certain signaling pathways, luciferase reporters have been generated that have a low intrinsic stability to quickly reflect transcriptional dynamics by adding a rapid degradation signal. As shown before, dual imaging using different luciferases is possible [28, 30, 31]. Moreover, bioluminescence imaging can be supplemented with other optical imaging (fluorescence) or other imaging techniques (multimodal imaging, see also [32]).

• 5. To compare the mammalian codon-optimized firefly luciferase 2 with the conventional firefly luciferase, stable luciferase 2 positive PC-3M-Pro4 prostate cancer cell lines were generated using a vector containing the enhanced luciferase gene under control of the CAGGS promotor (as described in [33]). Several clones were selected for further analysis and their luciferase activities were measured. Although variations occurred between the different clones, on average the PC-3M-Pro4 luciferase 2 (PC-3M-Pro4luc2) clones emitted 95 RLU/cell; this is on average a 202-fold (202 ± 95) increase compared to PC-3M-Pro4 cells with the previous luciferase generation (PC-3M-Pro4luc, 0.7 RLU/cell). We assessed the levels of bioluminescent signal of a titration series of PC-3M-Pro4luc vs. PC-3M-Pro4luc2 cells (10–50,000 cells). As shown in Fig. 2, lower amounts of PC-3M-Pro4luc2 cells compared to PC-3M-Pro4luc could be measured using an in vitro luciferase assay (Dual Luciferase Reporter Assay System, Promega). Even for the lower range of cells (10–1000 cells), an R ² was calculated of approximately 1, indicating that luciferase 2 activity can predict the amount of cells quite accurately (Fig. 2a, b for respectively high and low range of cells). In contrast, PC-3M-Pro4luc cells were not accurately measurable using a luciferase assay in these low amounts (R ² = 0.1). We subsequently compared the PC-3M-Pro4luc with the PC-3M-Pro4luc2 in a titration series implanted subcutaneously (Fig. 2c, d). The 10,000 and 1000 cell conditions were accurately and reproducibly measurable in all mice inoculated. In all mice (10/10), the 100 PC-3M-Pro4luc2 cell condition could be measured, whereas no measurable signal was detected in the 100 PC-3M-Pro4luc cells condition (0/10). In 1/10 even 20 PC-3M-Pro4luc2 cells could be detected subcutaneously using the IVIS Lumina. We next inoculated prostate cancer cells in the most frequent metastatic site in prostate cancer, the bone. As shown in Fig. 3, PC-3M-Pro4luc2 displayed significantly enhanced bioluminescent signal compared to PC-3M-Pro4luc (on average 130-fold increase) although tumor volume was similar (Fig. 3c, d).

  To compare the metastatic properties, the PC-3M-Pro4luc and PC-3M-Pro4luc2 cells were inoculated into the left cardiac ventricle. As shown in Fig. 4, BLI signal could already be detected after 24 h in the mice inoculated with PC-3M-Pro4luc2 cells. In contrast, BLI signal appeared 14 days after inoculation of PC-3M-Pro4luc cells. Total tumor burden of mice inoculated with PC-3M-Pro4luc2 cells increased in the first week, then decreased until day 10 followed by exponential growth thereafter. The presence of tumor cells has been validated by immunohistochemical staining (Goldner staining) showing a metastasis in the bone marrow (Fig. 4d, e). An example of a micrometastasis (measured after 48 h) in a tibia (Fig. 4f) displays possible arrest in the bone marrow vasculature. The number of metastases in the mice inoculated with PC-3M-Pro4luc2 cells was initially significantly higher, but after 14 days no significant differences were found. Most likely, these metastases are also present in the mice inoculated with the PC-3M-Pro4 luc cells, but remain below the detection limit of the IVIS Lumina. We further examined the tumor homing and early growth using the PC-3M-Pro4luc2 cell line. As shown in Fig. 5a, some metastases form but do not grow into overt metastases. For example, the metastasis in the left tibia grows into an overt metastasis, whereas the metastasis in the right tibia disappears and becomes undetectable after 10 days (with both IVIS Lumina measurements and immunohistochemistry). Moreover, the first few days many micrometastases can be detected in the distal spine of which none grow into overt metastasis, whereas some metastases in the lumbar spine do develop into overt metastases (Fig. 5b, c). This indicates that the prostate cancer cells home to these sites, but do not grow into overt metastasis. This model can therefore be used to study the kinetics of homing and metastasis of prostate cancer cells in real time.

• 6. Which cell line to use depends on the research question. Cell lines (stably) expressing reporter genes (either fluorescent or bioluminescent ) are needed for the optical imaging procedures. Otherwise caliper measurements should be performed supplemented with species-specific qPCR on several tissues that might contain metastases.

• 7. A significant problem for optical imaging using fluorescence is the limited transmission of the light through the tissues due to scattering and absorption of the light in the tissues. Autofluorescence from endogenous molecules, such as hemoglobin and cytochromes, causes significant background signals. Since autofluorescence of the tissue is less prominent and tissue penetration is improved in the far-red or near-infrared (NIR), fluorescent proteins in these regions are more suitable for deeper, noninvasive imaging of small animals (reviewed in [34]). For example, living colors© fruit fluorescent proteins (RFP variants) such as dTomato and mCherry are more suitable for in vivo experiments in small laboratory animals. In addition to autofluorescence, the limited depth of the fluorescence signaling impairs the feasible experiments. Combinations of high-speed intravital imaging using a multiphoton laser scanner microscope with the dorsal skinfold window model has increased the depth of the tissue that can be interrogated [35]. However, experiments are still restricted to the dorsal skinfold chamber or subcutaneous regions. It remains difficult to measure sensitively in bones. Compared to bioluminescence , relatively large tumors (~1 mm³) are needed to measure fluorescent signals in bone (reviewed in [24]).

• 8. The simplest approach to labeling cells for cell tracking is to add a fluorescent dye that is taken up by viable cells. However, because these dyes are diluted during each cell division, the signal per cell reduces over time, thus limiting some long-term experiments. However, this procedure can be useful in determining a cell population with limited cell division (e.g., stem cells) (reviewed in [36]).

• 9. In case of PC-3M-Pro4 cells and other established prostate cancer cell lines, Balb/c nu/nu are a suitable animal model. Other immune-deficient mice can also be utilized. For example, for experiments with the established prostate cancer cells VCaP or C4-2B cells, NOD SCID mice are preferred since these cells did not show tumor growth in Balb/c nu/nu mice. NSG mice are usually adequate for experimental metastasis procedures. Regarding the choice of the appropriate mouse model, it should be taken into consideration that light is transmitted more efficiently through the tissues of white or hairless mice because melanin absorbs substantial amounts of light and light is scattered more by dark fur.

• 10. Fat depositions surrounding the prostate and other organs are rarely found in young Balb/c nu/nu mice but more regularly in other mouse strains and older Balb/c nu/nu mice.

• 11. 4.5–5-week-old mice preferably since the amount of bone metastasis is highest since these mice have higher bone turnover compared to older mice.

• 12. The amount of cells inoculated depends on the experiment. When using drug efficacy studies, 10⁵ cells/100 μL can be inoculated. However, determining the effect of different subpopulations of cells can be performed with less cells. For example, inoculation of 10,000 cells of the ALDH^high stem/progenitor subpopulation resulted in tumor progression and metastasis [19].

• 13. Occasionally, the luciferin is misinoculated intraperitoneally. If no or less bioluminescence signal appears than expected because of previous measurements or a palpable tumor, you should reinject d-luciferin.

• 14. A problem of bioluminescence imaging can be overexposure of smaller foci by large signals. It is possible to cover the large focus with a piece of black carton or plexiglass to measure smaller foci. Measure mice separately when one mouse has a significantly higher tumor burden. At the end of the experiment, quickly resect the large tumors and subsequently remeasure the mouse to visualize potential additional (micrometastatic) tumors previously not visible. Another possibility is to isolate the organs presumably containing tumor cells and measure the organs ex vivo.

• 15. If the image is saturated, first the F-stop (diaphragm) can be adjusted. It is not necessary to remeasure the other animals, since changing the F-stop does not change the measured bioluminescent signal. The software corrects for the F-stop.

• 16. Histological analysis of the BLI signal remains pivotal, since environmental factors and therapeutic interferences may cause discrepancies between tumor burden and bioluminescence intensity. The intensity of the signal measured by in vivo imaging may depend on various factors, such as d-luciferin absorption through the peritoneum, cell membrane permeability, availability of co-factors, intracellular pH, and transparency of overlying tissue. In addition, because luciferases are oxygenases, the oxygen requirement may limit the use of luciferases as reporters in anaerobic environments, such as necrotic cores of large tumors. Similarly, the blood flow may not be sufficient to deliver the substrate to the tumor cells expressing luciferase.

Fig. 2

Comparison between Pro4luc and Pro4luc2 cells growing in vitro and in a subcutaneous in vivo model. Quantification of biochemically measured luciferase activity and its correlation with cell number with a cell number range of respectively 10–50,000 (a) and 10–1000 (b). The R ² coefficient was calculated with SPSS 17.0. Representative images of mice with subcutaneously inoculated PC-3M-Pro4luc or PC-3M-Pro4luc2 cells 3 h after inoculation (c) with the quantification in relative light units in (d). (open circles: Pro4luc; closed circles: PC-3M-Pro4luc2). Data are representative of three independent experiments (n = 4)

Fig. 3

Comparison between Pro4luc and Pro4luc2 cells growing intraosseous. Representative images at day 3 after inoculation of either PC-3M-Pro4luc or PC-3M-Pro4luc2 cells in the bone (a). Total tumor burden for the mice inoculated with either PC-3M-Pro4luc2 (closed circles) or PC-3M-Pro4lucvehicle (open circles). In the insert, the first 14 days of the experiment are shown (b). (open circles: PC-3M-Pro4luc; closed circles: PC-3M-Pro4luc2). Representative images of a Goldner staining showing respectively PC-3M-Pro4luc (c) and PC-3M-Pro4luc2 (d) tumor cells in the tibia. GP growth plate, T tumor, B bone. Scale bars represent 100 μm

Fig. 4

Comparison of Pro4luc and Pro4luc2 cells in an experimental metastasis model. Representative images of mice at day 1, 2, 7, 10, 14, 17, 21, and 24 after inoculation of either PC-3M-Pro4luc or PC-3M-Pro4luc2 cells (a). Total tumor burden for the mice inoculated with the respective cells. In the insert, the first 17 days of the experiment are shown (b). The number of metastases/mouse (c). (open circles: PC-3M-Pro4luc; closed circles: PC-3M-Pro4luc2). Bioluminescent images of mice at day 24 with RLU of the left tibia and corresponding images of a Goldner staining showing respectively mice inoculated with PC-3M-Pro4luc (d) and PC-3M-Pro4luc2 (e) cells. GP growth plate, T tumor, B bone. Scale bars represent 80 μm. Representative image of a mouse 48 h after inoculation of PC-3M-Pro4luc2 cells (f). Goldner staining of the left tibia of this mouse (J) with corresponding magnification. B trabecular bone, BM bone marrow, S sinusoids, T tumor, OC osteoclast, Ocy osteocyte, Ob lining osteoblast

Fig. 5

Kinetics of tumor growth. Representative images at respectively day 1, 2, 7, 10, and 14 after inoculation of PC-3M-Pro4luc2 cells (a). Representative images of mice 24 h after inoculation of PC-3M-Pro4luc2 cells with a lower threshold setting (b). Amount of metastases/mouse at different sites (c)