2.1 Cell culture

LNCaP cells stably expressing firefly luciferase (LNCaP-CMV-Luc) were a generous gift from Ronald Rodriguez (Johns Hopkins University). Cells were maintained under G418 selection in RPMI-1640 (GIBCO, 11875-093) with 10% FCS at 37°C in 5% CO₂. Cell line authentication and mycoplasma testing was completed through the Johns Hopkins Genetic Resources Core Facility JHU.

2.2 Subcutaneous tumor model

Animal studies were performed according to protocols approved by the Animal Care and Use Committee at Johns Hopkins University. Athymic (nu/nu) immunocompromised male mice were purchased commercially (Taconic Biosciences, NCRNU-M sp/sp). For subcutaneous tumor generation, 5 × 10⁶ LNCaP-CMV-Luc cells were suspended in a 1:1 suspension with Matrigel (Corning, 354248) and inoculated subcutaneously in the right dorsal flank in a volume < 200 μl/animal. Tumor volume was monitored by slide caliper (Mituoyo) measurements of length, width, and height, and then using the equation: 0.52 × L × W × H (in mm³).

2.3 Intratibial tumor model

Animals were anesthetized with ketamine: xylazine: acepromazine (100:20:3 mg/kg). The knee was shaved and sterilized. A para-patellar incision (0.5 cm long) was created to expose the injection area. A 25G Hamilton syringe needle was inserted through the cortex of the anterior tuberosity of the tibia with a rotating “drill−like” movement to minimize cortical fracture. After the bone cortex was traversed, the needle was further inserted approximately 3 mm down to the diaphysis of the tibia to create a hollow core. 1 × 10⁶ LNCaP-CMV-Luc cells were prepared in 10−15 µl of 1× phosphate-buffered saline (PBS) and injected slowly. Bone wax (Surgical Specialties Corporation, 903) was applied to seal the drilled area and prevent exodus of implanted cells. The incision was closed with sterile 4-0 chromic suture (Ethicon Inc., 6543478) on skin. Buprenorphine was administered for pain management.

2.4 Histology and microCT

Tibias were harvested and fixed in 10% formalin for 48 h. Fixed tibias were scanned using micro-CT (nanoScan PET/CT; Mediso, Budapest) and subsequently paraffin-embedded, sectioned (5 μm thick) and stained with hematoxylin and eosin (H&E) for histologic analyses.

2.5 Bioluminescent tumor imaging

Bioluminescence imaging was performed using the IVIS Spectrum imaging system (Xenogen, Caliper Sciences). Before imaging, each mouse was intraperitoneally injected with 3 mg of luciferin (Gold Biotechnology, 115144-35-9). Animals were anesthetized by the open-drop isoflurane exposure method and maintained at 1.5%–2% isoflurane anesthesia during imaging. All images were analyzed using Living Image 4.3 software under the same intensity scale to normalize across time. After normalization, a region of interest (ROI) was generated and applied to all tumors and average radiance (p/s/cm²/sr) calculated.

2.6 ctDNA extraction and quantification

Approximately 35 μl of tail vein blood was drawn into a marked heparinized micro-capillary tube (Thermo Fischer Scientific, 22-362-566) fitted with dropper cap (Drummond Microcaps, 101-000-0100). The collected blood was transferred to a 1.5 ml tube containing 200 μl of 4.5 mmol/L EDTA in PBS, centrifuged at 1500 × g for 15 min, and plasma supernatant was collected and stored at −80°C. DNA was extracted by QIAamp DNA Micro Kit (Qiagen, 56304) and eluted in 50 μl DNAse free water. Human LINE-1 and Alu ctDNA was quantified using iTaq™ Universal SYBR® Green Supermix (BioRad, cat no.1725121) and primers Human_LINE1_Fwd (TCACTCAAAGCCGCTCAACTAC) and Human_LINE1_Rev (TCTGCCTTCATTTCGTTATGTACC) as reported by Rago et al.,² or primers Human_Alu_115_Fwd (GTCAGGAGATCGAGACCATCCT) and Human_Alu115_Rev (AGTGGCGCAATCTCGGC) as derived from Asangani et al.²⁰ Samples with Threshold cycle (Ct) values above background controls were excluded. Serially diluted LNCaP-CMV-Luc DNA was spiked in the background of nonhuman DNA (salmon sperm, Thermo Fisher Scientific, cat no.15632011) and applied as a reference standard curve.

2.7 CT guided focal ionizing radiation (IR) treatment of animals

Ionizing radiation (IR) was delivered using the Small Animal Radiation Research Platform (SARRP) (Xstrahl). Mice were anesthetized using 1.5%–2% isoflurane gas and then immobilized on the SARRP's rotary stage. The SARRP was equipped with cone-beam computed tomography (CBCT) and a treatment planning system, which were used to image each animal to create a highly targeted and conformal IR delivery. For the subcutaneous tumor group, an AP (anterior to posterior) beam of 10 × 10 mm at a variable depth (due to tumor size) was used. For the tibia group, an AP beam of the same size with the radiation isocenter on the bone was used. The energy of the radiation beams used was 220 kVp and 13 mA. The slice gap was one degree and the resolution was 0.2 mm. Animals were treated with a nonablative single fraction of 6 Gy, 28 days postimplantation of LNCaP-CMV-Luc cells. Reconstruction of images based on data was done by software built into the machine. Tumors were identified as highly contrasted masses present in animals but absent in control (noninjected mice).

2.8 Statistical analysis

Each study cohort comprised N = 5 animals. Tumor growth and therapeutic response were recorded for animals that developed tumors, with 5/5 subcutaneous nonirradiated, 5/5 subcutaneous irradiated, 4/5 intratibial nonirradiated, and 3/5 intratibial irradiated cohorts developing tumors. Spearman correlation analysis was performed using GraphPad Prism 8.0 (GraphPad). p < .05 was considered statistically significant for all the interpretations. All tumor measurements are provided in Table S1 and Figures S1 and S2.

3.1 Models for comparative analysis of tumor burden by concurrent ctDNA, bioluminescence, and caliper measurement

LNCaP cells, engineered to stably express firefly luciferase, were used to generate an intratibial human tumor xenograft model in athymic nude mice. We applied X-ray, microCT, bioluminescence imaging, and tumor histology to optimize tumor engraftment and verify tumor retention within the marrow space (Figure 1A-C). We found that bone wax was important to seal the injection site and prevent tumor passage out of the bone. Once the model was optimized, we developed cohorts of mice bearing subcutaneous or intratibial xenograft tumors for comparative measurement of tumor growth by slide caliper measurement (subcutaneous only), bioluminescence imaging, and quantitative real-time PCR (qPCR) of circulating human tumor DNA (ctDNA) using LINE-1 or Alu transposable element amplicons. We completed all tumor measurements and plasma sampling within a single day to allow direct comparison of each measurement approach in the same animal (see Figure 1D for study schematic). A subset of animals received a single, non-curative dose of external beam IR (6 Gy) to the tumor region using the CT-guided SARRP to compare the different measurement approaches in the context of therapeutic response and tumor regrowth.

3.2 ctDNA level, bioluminescence, and caliper measurement are significantly correlated in subcutaneous prostate tumor xenografts

Following subcutaneous tumor engraftment, we concurrently measured tumor burden on multiple days by slide caliper measurement, bioluminescence, and LINE-1 and Alu ctDNA level. The data from each measurement was log2 transformed for comparative graphing analysis. All three measurements demonstrate a concordant increase in tumor size and signal over time (Figure 2A-C). We performed Spearman correlation analysis to compare the relationships between each measurement at every time point. Plasma concentrations of LINE-1 and Alu ctDNA were strongly and significantly correlated with tumor volume (Figure 2D), with Spearman r = .9377 for LINE-1 ctDNA with tumor volume, and Spearman r = .9477 for Alu with tumor volume (p < .0001 for both amplicons, independently). Levels of ctDNA and bioluminescent radiance were also significantly correlated (Figure 2E), with calculated Spearman r = .6562 for LINE-1 and Spearman r = .5954 for Alu (p < .002 for both amplicons, independently). Bioluminescent radiance was also significantly correlated with tumor volume (Figure 2F), with calculated Spearman r = .6646 (p = .0003). These results validate human LINE-1 and Alu ctDNA quantification as methods for assessing subcutaneous human PCa xenograft tumor burden in athymic nude mice.

3.3 ctDNA level and bioluminescent radiance are significantly correlated in intratibial prostate tumor xenografts

To assess the accuracy of ctDNA in the measurement of inaccessible tumors in bone, we applied intratibial LNCaP-MLuc xenografts. Following engraftment, we measured ctDNA and bioluminescence for multiple days over a 2-month experiment. Both ctDNA level and bioluminescent radiance concurrently increased in signal over time (Figure 3A,B). Plasma concentrations of LINE-1 and Alu ctDNA were strongly and significantly correlated with tumor bioluminescence (Figure 3C), with a calculated Spearman r = .816 for LINE-1 ctDNA and bioluminescence, and Spearman r = .813 for Alu ctDNA and bioluminescence (p < .0001 for both amplicons, independently). These results further validate the use of LINE-1 and Alu ctDNA as noninvasive biomarkers for assessing human PCa xenograft tumor burden within the bone marrow of athymic nude mouse models.

3.4 Evaluation of ctDNA, bioluminescence, and tumor volume in irradiated subcutaneous and intratibial prostate tumor xenografts

We then assessed the ability of ctDNA to measure tumor response to a single nonablative dose of IR therapy by comparison to bioluminescence and caliper (subcutaneous only) tumor measurement for several days over a 2-month experiment. We applied external beam radiation therapy because it is a standard treatment for the palliation of bone pain caused by PCa bone metastases and a developing approach for the treatment of oligometastatic PCa. For this model, LNCaP-MLuc cells were engrafted subcutaneously, or within the tibia, and tumor burden was concurrently measured by all methods before radiation, on the day of radiation (single dose, 6 Gy), and several days the following radiation. The average measurements at each time point are shown in Figure 4, and individual measurements of each tumor are provided in Figure S3. All three measurement approaches produced comparable growth and therapeutic response curves for subcutaneous tumors, although transient differences in signal were apparent for a few days following IR (Figure 4A-C). Caliper measurement reported stabilization of tumor volume in all animals for several days following IR (Figures 4A and S3A, Days 28–36), while bioluminescence signal markedly decreased and the recovered after IR (Figures 4B and S3B, Days 28–36). The ctDNA levels varied for a brief time following IR, with signal increasing in some animals and decreasing in others (Figures 4C and S3C, Days 28–36). In the intratibial model, both bioluminescence and ctDNA signals briefly decreased after radiation, and then recovered and increased after several days (Figures 4D,E and S3D,E).

To compare the concordance of ctDNA, bioluminescence and tumor volume over the course of the irradiated tumor model experiments, we performed Spearman correlation analysis. The correlation between ctDNA and tumor volume was statistically significant in irradiated subcutaneous tumors (Figure 5A), with Spearman r = .6360 for LINE-1 and caliper measurement, and Spearman r = .4411 for Alu and caliper measurement (p < .02 for both amplicons, independently). However, the correlation between bioluminescence and caliper measurement (Figure 5B), and bioluminescence and ctDNA (Figure 5C), were not statistically significant in irradiated subcutaneous tumor models, possibly due to the transient differences in signal response following IR. The correlation between ctDNA and bioluminescence in intratibial tumors was statistically significant for LINE-1 analyses, with Spearman r = .55 (p = .018), but the correlation between Alu ctDNA and bioluminescence did not reach statistical significance (Figure 5D). The absence of a statistically significant correlation may be due to the transient, differential responses of the different measurement signals following radiation.