Brachytherapy using injectable seeds that are self assembled from genetically encoded polypeptides in situ

Abstract

Brachytherapy is a common clinical technique involving implantation of sealed radioactive "seeds" within a tumor to selectively irradiate the tumor mass while minimizing systemic toxicity. To mitigate the disadvantages associated with complex surgical implantation and subsequent device removal procedures, we have developed an alternative approach using a genetically-encoded peptide polymer solution composed of a thermally responsive elastin-like polypeptide (ELP) radiolabeled with ¹³¹I that self-assembles into radionuclide seeds upon intratumoral (i.t.) injection. The formation of these non-toxic and biodegradable polymer seeds led to prolonged intratumoral retention (~85% ID/tumor 7 days post-injection) of the radionuclide, elicited a tumor growth delay in 100% of the tumors in two human xenografts (FaDu and PC-3), and cured more than 67% of tumor-bearing animals after a single administration of labeled ELP. These results suggest that in situ self-assembly of biodegradable and injectable radionuclide-containing polypeptide seeds could be a promising therapeutic alternative to conventional brachytherapy.

Introduction

Brachytherapy is a clinical technique that is typically used in the treatment of prostate cancer and other malignancies to selectively expose a tumor to large doses of beta and gamma radiation via locally implanted “seeds” (1). The strategic placement of these seeds throughout the tumor mass can provide prolonged irradiation to the tumor while limiting the radiation burden on healthy tissues. Although the success of current approaches to prostate brachytherapy is exemplified by excellent disease-free survival (2–4), its implementation is plagued by a number of significant limitations, including the need for general anesthesia, complicated placement procedures, seed migration to other organs, and the need for post-treatment excision for seed removal (5). We speculated that many of these limitations might be overcome by engineering a biodegradable material that a) can form injectable depots with a life span in vivo that is compatible with the half-life of the radionuclide used for treatment; and b) degrades into non-toxic products that are resorbed by the body after the radioactivity has decayed to a level that no longer poses a threat of systemic toxicity, thereby obviating the need for surgical implantation and re-excision. Although various polymers have been designed and evaluated with these requirements in mind, issues such as toxicity, biological incompatibility, a lack of injectability and the less ability to scale-up manufacturing continue to hinder their clinical implementation (6–9).

Elastin-like polypeptides (ELPs) are a unique class of genetically engineered peptide polymers that have several attractive properties for brachytherapy. First, ELPs undergo phase separation in response to small temperature changes. ELPs remain highly soluble in aqueous solution below a characteristic transition temperature (T_t), but within seconds, desolvate into a hydrophobic coacervate above the T_t. This temperature is tunable by adjusting the amino acid composition, molecular weight (MW), and concentration of the ELP (10–12). The ability to tune the behavior at the molecular level allows for the facile design of ELPs that can serve as an injectable depot for radioactivity or other drugs. Second, ELPs are biodegradable, biocompatible, and their design at the genetic level enables facile addition of functional groups for radiolabeling (13–15) and provides exquisite control over the composition and MW of the peptide. This versatility facilitates the design of ELPs that are soluble at room temperature but rapidly collapse into a viscous coacervate upon warming to body temperature. This feature is attractive for the formation of an injectable depot of covalently conjugated radionuclide for brachytherapy, as it avoids the need for surgical placement of the radioactive depot. In a previous study, the ELP phase transition temperature was adjusted to drive ELP coacervation immediately following tumor infusion. The formation of a coacervate phase within the tumor increased the intratumoral retention of the radiotherapeutic, thereby enhancing the antitumor efficacy compared with a thermally insensitive control ELP. However, this [¹³¹I]ELP depot was not able to induce complete tumor regression (16).

Herein, we engineered an easily injectable radiolabeled ELP that thermally self-assembles into highly stable, therapeutic radionuclide seed-like depots in situ upon i.t. administration and optimized their in vivo performance. To optimize the therapeutic efficacy of this new modality, we investigated the effect of a range of parameters to modulate the physical properties of the depot, including the ELP MW, concentration, composition, and the tyrosine-rich peptide appended to the carboxy-terminus for radiolabeling. By designing a diverse set of injectable ELPs, we show that the temporal stability of the radioactive depot in vivo can be tuned by controlling the phase transition temperature of the radionuclide carrier and present evidence of the therapeutic efficacy of this engineered thermally responsive biomaterial. The optimized seed-like depot displays prolonged intratumoral retention and exhibits potent antitumor efficacy, and eventually degrades into nontoxic peptides.

Materials and Methods

ELP design and synthesis

A thermally responsive ELP composed of tandem repeats of the pentapeptide motif VPGVG was designed to transition below 30 °C to enable immediate coacervation upon intratumoral infusion within the concentration range of 62.5 -1000 µM. To evaluate the effect of MW on tumor retention, synthetic genes encoding three ELPs spanning a range of MWs were synthesized using PRe-RDL (17). These three ELPs, consisting of 60 pentapeptides (25 kDa), 120 pentapeptides (50 kDa) and 240 pentapeptides (100 kDa), are referred to as ELP₆₀, ELP₁₂₀ and ELP₂₄₀, respectively. ELP₆₀ and ELP₂₄₀ included a single tyrosine residue at the C-terminus, and are referred to as ELP₆₀-Tyr₁, and ELP₂₄₀-Tyr₁. ELP₁₂₀ was modified to include 1, 4, or 7 tyrosine residues at the C-terminus, and these are referred to as ELP₁₂₀-Tyr₁, ELP₁₂₀-Tyr₄, and ELP₁₂₀-Tyr₇, respectively. Synthetic genes encoding these ELPs were inserted into pET-24a (+) expression vectors and transformed into BL21™ E. coli chemically competent cells (Edge BioSystems; Gaithersburg, MD). Transformed cells were used to inoculate one liter of TB-Dry media (Mo Bio, Carlsbad, CA) in a 4-liter flask, which was supplemented with 45 µg/mL kanamycin and incubated for 24 h at 37 °C and 200 rpm. ELP was purified using Inverse Transition Cycling (ITC) (18), a non-chromatographic method that exploits the ELP thermal transition to sequentially remove soluble and insoluble impurities through centrifugation. ITC also enables the recovery of ELP at high yields with endotoxin levels that are below the FDA required limit of 5 EU/dose (1 dose = 1 mg). The ELP thermal properties were characterized by monitoring the optical density (OD) of the ELP solution at 350 nm as a function of temperature (1°C/min) on a temperature-controlled UV-Vis spectrophotometer (Cary 300 Bio; Varian instruments, Palo Alto, CA).

Radioiodination of ELP

ELPs with carboxy-terminal tyrosine residues were labeled with either ¹²⁵I or ¹³¹I (PerkinElmer, Boston, MA) using the IODO-Gen kit (GE Healthcare; Piscataway, NJ). Briefly, 100 µL of ELP at a specified concentration in PBS was applied to an IODO-Gen pre-coated tube containing 2 mCi Na ¹²⁵I (for tumor retention and biodistribution studies) or 20 mCi Na ¹³¹I (for radiotherapy studies), and the reaction product was purified by size-exclusion chromatography with a PD-10 column (GE Healthcare; Piscataway, NJ). The radioactivity was quantified by counting an aliquot o n a γ-counter (LKB-Wallac; Turku, Finland). The concentration of iodinated ELP was measured by UV-Vis spectrophotometry (Thermo Scientific; Waltham, MA) at a wavelength of 280 nm. The final radioactivity and molar concentration of ELP was adjusted by mixing labeled and unlabeled ELPs to achieve the desired injection concentration and dose. The radioactive dose of [¹²⁵I]ELPs for tumor retention studies was 10 µCi ¹²⁵I / 40 µL of ELP with a specific radioactivity of 6 – 12 µCi/mg, while for biodistribution studies the dose was 3 µCi ¹²⁵I / 40 µL of ELP at 1000 µM with a specific radioactivity of 1 – 2 µCi/mg. To assess therapeutic efficacy, we used [¹³¹I]ELPs at 1000–2000 µCi ¹³¹I / 40 µL of 1000 µM ELP with a specific radioactivity of 7500 – 15,000 µCi/mg.

Animal implantation of tumor subcutaneous xenografts

Female nude mice (Balb/c, nu/nu) with an average body weight of 20 g were provided by the Duke Comprehensive Cancer Facility Isolated Facility for all in vivo experiments, including ELP tumor retention, biodistribution, and [¹³¹I]ELP radiotherapy. The mice were housed in isolated caging with sterile rodent food, acidified water ad libitum, and a 12 h light/dark cycle. The FaDu head and neck squamous cell carcinoma cell line was obtained from the American Type Culture Collection and maintained by Duke University Cell Culture Facility (DUCCF). PC-3M-luc2, which is a derivative of the PC-3 cell line, is a luciferase-expressing metastatic prostate cancer cell line that was stably transfected with firefly luciferase gene (luc2) (Caliper Life Science; Hopkinton, MA). Both cell lines were authenticated by STR-DNA technology (RADIL) by DUCCF, and Caliper Life Science respectively, when we bought them and started cell culture for this study in our lab. The two tumor cell lines were cultured as monolayers in tissue culture flasks. The culture medium contained minimal essential medium (MEM) supplemented with 10% heat-inactivated FBS and a 1% antibiotic/antimycotic solution (Invitrogen, Carlsbad, CA). The cultures were maintained at 37 °C under 5% CO₂. Leg tumor xenografts were established by s.c. inoculation of 1 × 10⁶ cells in 30 µL of MEM medium in the right lower leg of each mouse. Tumors were allowed to grow for 7–9 (FaDu) or 11–13 (PC-3) days before treatment until they reached a size of 150 ± 20 mm³. Mice were monitored for general well-being, weight, and tumor volume. All animal experiments were performed in accordance with the Duke University Institutional Animal Care and Use Committee regulations.

Tumor retention of ELP conjugates

To investigate the effect of the injected concentration, MW, and tyrosine content on the tumor retention of ELP, three retention experiments were performed, varying: (1) the injection concentration of ELP₁₂₀-Tyr₁ (62.5, 125, 250, 500, and 1000 µM); (2) MW of the ELPs (ELP₆₀- Tyr₁, 25 kDa; ELP₁₂₀-Tyr₁, 50 kDa; and ELP₂₄₀-Tyr₁, 100 kDa); and (3) the tyrosine content at the C-terminus of ELP₁₂₀ at an injection concentration of 250 µM (ELP₁₂₀-Tyr₁, ELP₁₂₀-Tyr₄, and ELP₁₂₀-Tyr₇). Each ELP was infused into the tumor using a syringe pump (Harvard Apparatus; Holliston, MA) at a rate of 120 µL/min in a volume of 40 µL into a 150 mm³ tumor. The radioactivity retained in the mice was monitored with a dose calibrator (Biodex Medical Systems; Shirley, NY) at 0, 24, 48, 72, 96, and 168 h.

Rheological characterization

The ELP viscosity in PBS was measured on an AR G2 rheometer (TA Instruments, New Castle, DE) with a concentric cylinder. An ELP solution in PBS was loaded onto a preheated cylinder (35 °C), the samples were compressed to a height of 55 µm and allowed to equilibrate for 2 min at 35°C. Steady shear measurements were performed over a range of shear rate between ~10⁻³ and ~10³ s⁻¹.

Dynamic Light Scattering

Dynamic light scattering (DLS) was performed on ELP₁₂₀-Tyr₁, ELP₁₂₀-Tyr₄ and ELP₁₂₀-Tyr₇ using a thermally controlled Wyatt Plate Reader (Wyatt Technology, Santa Barbara, CA) to determine the hydrodynamic radius (R_h). The samples were diluted to 50 µM in PBS, filtered through a 0.1 µm Anotop syringe filter (Whatman), and 35 µL of each solution was placed into a well of a 384-well plate (Corning, Corning, NY). Small drops of mineral oil were added to the top of each well to prevent evaporation. Each sample was analyzed with eighteen 10 second acquisitions at 15 °C, which was below the transition temperature for these ELP. The resulting data were fit using a Rayleigh sphere model with a regularization algorithm. Populations comprising less than 3% of the total mass were excluded from analysis.

Tumor Retention and Therapeutic Efficacy of [¹³¹I]ELPs

The effect of increased tyrosine content on the in vivo tumor retention and antitumor efficacy of [¹³¹I]ELP was examined in a s.c. FaDu tumor model by i.t. administration of the following ELP solutions: [¹³¹I]ELP₁₂₀-Tyr₁, [¹³¹I]ELP₁₂₀-Tyr₄, and [¹³¹I]ELP₁₂₀-Tyr₇ at an ELP concentration of 1000 µM and a total radioactivity dose of 2000 µCi into a 150 mm³ tumor. The in vivo retention and antitumor efficacy of [¹³¹I]ELP₁₂₀-Tyr₇ was also examined in subcutaneous human prostate PC-3M-luc2 xenografts in nude mice. Both tumors were compared with an unlabeled ELP control, also administered at 1000 µM. To prevent thyroid accumulation of free radioiodine, 1% potassium iodide was added to the drinking water of all of the animals one week before treatment and continued till the end of the radiotherapy experiment. The mice were monitored daily for tumor radioactivity, tumor volume, body weight (BW), and survival during the first week, every other day during the first month, and twice a week thereafter. The tumor radioactivity profile was monitored with dose calibrator after i.t. injection of [¹³¹I]ELP until the level fell below 5% of the administered level in each mouse. The tumor volume was determined by the equation: volume = (width)² × length × π/6. The mean tumor growth over time per group was calculated and expressed as the relative tumor volume compared to day 0 (i.e., fold change). Median values were used instead of the mean because some tumors regressed completely. Animals were humanely killed once any of the following end-points were reached in the course of the radiotherapy: the tumor size reached 5 times the initial tumor volume, the tumor volume reached 1 cm³, the body weight dropped below 85% of the body weight or 60 days post-treatment (the end of the experiment).

Safety evaluation of ELPs

The safety evaluation of the ELP radionuclide depot was examined through two experiments: 1) the radioiodine toxicity during radiotherapy was evaluated by monitoring clinical signs of health and changes in body weight; and 2) the organ and tissue toxicity was examined by measuring the biodistribution of ¹²⁵I]ELP₁₂₀-Tyr₇ after i.t. administration in nude mice bearing FaDu tumors. Six mice were sacrificed at predetermined time points, their organs (blood, thyroid, lungs, heart, liver, spleen, kidneys, and tumor) were collected, and the radioactivity levels were measured with a γ-counter. The accumulated radioactivity was presented as injected dose per gram (%ID/g).

Histological studies and tumor autoradiography

On days 0, 1, 2, 3, and 60 after i.t. infusion of [¹³¹I]ELP, the animals were killed, and the tumor tissue was collected and fixed in 10% formalin. The formalin-fixed FaDu and PC-3 tumor tissue was then embedded in paraffin, sectioned (5 µm), and stained with hematoxylin-eosin. Tumor necrosis in both treatment and control groups was analyzed by light microscopy. Unstained tumor sections from the same set of samples were exposed to a storage phosphor screen (PerkinElmer, Boston, MA) for 4–5 days, which was then scanned with a Cyclone™ storage Phosphor (PerkinElmer, Boston, MA)

Statistical methods

Tumor retention data of [¹²⁵I]ELP and tumor growth data for [¹³¹I]ELP were analyzed by one-factor ANOVA analysis stratified by treatment group, which was followed by the Bonferroni t-test. P < 0.05 was considered statistically significant in both cases. Statistical differences between survival rates of the animals were determined by Kaplan-Meier analysis. The cumulative radioactivity and half-life of tumor radioactivity were analyzed with PK Solution 2.0 (Summit Research Services, Montose, CO, USA).

Results

Modulation of the phase transition behavior enables prolonged intratumoral retention

To promote intratumoral retention, each ELP was designed to form highly localized insoluble depots —analogous to seeds— at the temperature of subcutaneous tumors in anesthetized mice (31.7 °C). The T_t of each ELP satisfies this criterion under all examined conditions, including changes in concentration, MW, and tyrosine content (supplemental Table 1). The injected ELP is expected to remain assembled in a localized seed until its degradation and alternatively clearance by diffusion.

To determine the relationship between tumor retention and the physical properties of the ELP, concentration, MW, and tyrosine content were systematically varied. First, the effect of concentration was examined on the in vitro T_t and in vivo tumor retention of ELP₁₂₀ – Tyr₁ over a series of two-fold dilutions (1000 to 62.5 µM). Although the T_t decreased 3 °C over this 16-fold increase in concentration, the in vivo tumor retention 7 days following intratumoral infusion exhibited a 5-fold increase from 16% at 62.5 µM to 85% at 1000 µM (p < 0.01) (supplemental Table 1). These results indicate that higher concentrations could provide a significant increase in antitumor efficacy by increasing the tumor retention time of radioiodine.

Second, the effect of the ELP MW on the in vitro T_t and in vivo tumor retention of the ELP-seeds was examined using ELP₆₀ – Tyr₁ (25 kDa), ELP₁₂₀ – Tyr₁ (50 kDa), and ELP₂₄₀ – Tyr₁ (100 kDa). The ELP T_t exhibited a 4 °C decrease as the MW increased from 25 kDa to 50 kDa; this effect also translated to higher tumor retention, which demonstrated a 14-fold increase 7 days following infusion (p < 0.01). However, further increasing ELP MW to 100 kDa (ELP₂₄₀ – Tyr₁) only reduced the T_t by an additional 0.5 °C and failed to generate a significant increase in tumor retention from ELP₁₂₀ – Tyr₁ to ELP₂₄₀ – Tyr₁ (supplemental Table 1). These results also suggest that the ELP T_t plays an important role in tumor retention. Thus, ELP₁₂₀ – Tyr₁ (50 kDa) was selected for further characterization because it provided prolonged tumor retention while reducing the amount of ELP that must be administered.

Finally, we examined the effect of the number of tyrosine residues in the ELP on its tumor retention. It is generally understood that the ELP phase transition depends on the hydrophobicity of the sequence; as the hydrophobicity of the ELP increases, the T_t decreases (19). However, instead of altering the ELP sequence, a hydrophobic tyrosine-rich peptide was appended to the C-terminus to provide reactive sites for radioiodination as well as to decrease the T_t. As expected, the T_t of the ELPs decreased as a function of tyrosine content: ELP₁₂₀-Tyr₁ (24.97 °C), ELP₁₂₀-Tyr₄ (23.22 °C), and ELP₁₂₀-Tyr₇ (20.57 °C) all followed this trend. This reduction in the T_t caused a significant increase in the tumor retention of the seed, with each addition of 3 tyrosine residues inducing a 1.5 to 2-fold increase in tumor retention and a 2.5 °C reduction in the T_t (p < 0.01) (supplemental Table 1). Correlation analysis of the tumor retention experiments demonstrated that each of the three factors plays a critical role in the T_t: ELP concentration (r² = 0.983), ELP MW (r² = 0.998), and tyrosine content (r² = 0.961) (Fig. 1A).

Figure 1. ELP tumor retention is dependent on the transition temperature.

(A) In vivo tumor retention of ELPs as a function of ΔT_t. The ΔT_t is the difference between the tumor temperature (31.7 °C) in anesthetized mice and the ELP’s T_t. Each point in the figure is the tumor retention (%ID/tumor) 7 days after intratumoral infusion of each [¹²⁵I]ELP; (B) ELP viscosity after thermal phase transition as a function of ΔT_t. (C) Hydrodynamic radius (R_h) of ELPs at a temperature below its infusion temperature.

The tyrosine content also affected ELP viscosity; ELPs with greater numbers of tyrosine residues exhibited a higher viscosity than those with fewer residues (Fig 1B). This behavior may be partially explained by the presence of 36.9 ± 12.4 nm micelles in the ELP₁₂₀-Tyr₇ sample. Unlike ELP₁₂₀-Tyr₁ and ELP₁₂₀-Tyr₄ samples that solely consist of soluble unimers (R_h ≈ 6 nm), the ELP₁₂₀-Tyr₇ population was equally distributed between a unimeric state and a nanoparticle state as seen by DLS (Fig 1C). This suggests that when the nano-aggregates are heated to body temperature, the ELP coacervate that is formed contains a significant fraction of aggregated micelles, leading to a different internal microstructure than the other two ELPs, thereby affecting physicochemical properties such as its viscosity. These data are clear evidence of the importance of fine tuning the molecular architechture of ELPs and illustrate the many tools available for easily modulating intratumoral retention.

Long-lasting radiolabeled seeds lead to complete tumor regression

Tumor retention studies demonstrate that the 50 kDa ELP (ELP₁₂₀) administered at 1000 µM promotes the formation of long-lasting seeds. Furthermore, increased tyrosine content was also shown to enhance tumor retention and thus the cumulative tumor exposure to radioactivity (supplemental Table 1). Hence, to evaluate whether increased tumor retention translates into a therapeutic advantage, we investigated the tumor retention and antitumor effects of [¹³¹I]ELP₁₂₀ containing 1 (ELP₁₂₀-Tyr₁), 4 (ELP₁₂₀-Tyr₄), and 7 tyrosines (ELP₁₂₀-Tyr₇) per molecule at an injection concentration of 1000 µM. Figure 2A demonstrates that the temperature-triggered formation of an ELP seed containing 7 tyrosines most effectively retained the radioactive payload within the tumor mass. Correspondingly, the antitumor efficacy and survival time of the mice correlated with tumor retention, with ELP₁₂₀-Tyr₇ displaying the best therapeutic outcome, followed by ELP₁₂₀-Tyr₄ and ELP₁₂₀-Tyr₁ (Figure 2B–C). The ELP₁₂₀-Tyr₇ seed exhibited much better antitumor efficacy than either ELP₁₂₀-Tyr₁ or ELP₁₂₀-Tyr₄ and displayed the longest median survival time (60 days versus 39 and 17 days, respectively). In t he ELP₁₂₀-Tyr₇ treatment group, 10/12 mice survived until the end of the study (60 days), and 8/12 tumors showed complete regression. In contrast, all of the animals in the control ELP group, 6/7 mice treated with ELP₁₂₀-Tyr₁, and 6/8 mice treated with ELP₁₂₀-Tyr₄ were euthanized before the end of the study due to either excessive tumor size (> 1 cm³) (Table 1) or body weight loss (> 15 %) due to tumor burden. Histological examination illustrated that this radiation damage manifested as an acute inflammatory reaction (infiltration of white blood cells and congestion of blood vessels) in the tumor tissue at 24 h (Fig 2E) and the initiation of cell death at 48 h (Fig 2F) following administration. By 72 h, the tumor cell outline had disappeared (Figure 2G). However, tumor shrinkage was not observed until 5 days post-treatment (Fig 2B).

Figure 2. Tumor retention and antitumor efficacy of [¹³¹I]ELP in mice bearing FaDu tumor xenografts.

(A) Comparison of tumor retention of radioactivity, (B) tumor growth post-treatment, and (C) Kaplan-Meier plots of survival of mice treated with three different [¹³¹I]ELPs containing 1, 4 or 7 C-terminal tyrosine residues. Data represent mean ± SEM (n = 7–12). D, E, F, and G show HE stained tumor tissue sections obtained 0, 24, 48 and 72 h, respectively, after intratumoral infusion of [¹³¹I]ELP₁₂₀-Tyr₇.

Table 1.

Comparison of tumor retention of radioactivity and antitumor efficacy of the optimized [¹³¹I]ELP-depot in FaDu and PC-3 tumor xenografts in mice.

Parameter	FaDu				PC-3
	Unlabeled ELP120-Tyr4	[131I]ELP120 -Tyr1	[131I]ELP120 -Tyr4	[131I]ELP120- Tyr7	Unlabeled ELP120-Tyr7	[131I]ELP120 -Tyr7
Number of mice	11	7	8	12	10	13
Cumulative activity (day·µCi/tumor)	0	3597	8636	17086	0	15694
Half-life of radioactivity (day)	NA	3.2	5.27	7.08	NA	6.42
Growth rate at day 11 (mm3/day)	85	17	5	−5	53	−1.19
Median survival (day)	8	17	39	60	30	60
Complete regression	0/11	0/7	2/8	8/12	0/10	10/13
Survived at day 60	0 /11	1/7	2/8	10/12	0/10	13/13

The antitumor efficacy of these ELP seeds was also examined in PC-3 human prostate carcinoma xenografts. Both the tumor retention and the antitumor efficacy results for the PC-3 xenografts (Fig 3A–C) were comparable to those observed in the FaDu line (Table 1). In both tumor lines, [¹³¹I]ELP₁₂₀-Tyr₇ promoted prolonged tumor retention of ¹³¹I activity as measured by the area under the curve (cumulative radioactivity), half-life of tumor radioactivity (Table 1.) and suppressed tumor growth. Moreover, this treatment significantly prolonged the survival of mice bearing PC-3; whereas 10/10 of the control mice with PC-3 tumors died within 35 days, 13/13 of the treated mice survived until the end of the 60-d study. Histological examination of PC-3 tumors from these long-term survivors demonstrated that the ELP seeds eliminated all of the tumor cells remaining at the tumor site (Fig 3G, I) whereas the control ELPs permitted tumor growth and fibrosis (Fig 3F, H). These results demonstrate that this locally infused, self-assembled seed has significant therapeutic effect against two different human tumor lines.

Figure 3. Tumor retention and antitumor efficacy of [¹³¹I]ELP in mice bearing PC-3 tumor xenografts.

(A) Radioactivity retention, (B) antitumor efficacy, and (C) survival following intratumoral infusion o f [¹³¹I]ELP₁₂₀-Tyr₇. Data represent mean ± SEM (n = 10–13). Photographs of PC-3 tumors (D) 23 days after treatment with unlabeled ELP (E) 60 days after treatment with [¹³¹I]ELP₁₂₀-Tyr₇. H&E stained tumor tissue sections (F, H) obtained 23 days after treatment with unlabeled ELP or (G, I) 60 days after treatment with [¹³¹I]ELP₁₂₀-Tyr₇ (D).

The biodegradable radioactive seeds show minimal toxicity in mice

To address the concern that the biodegradability of these seeds and egress of labeled catabolites from tumor could lead to excessive radiation exposure of normal tissues, we monitored the overall health of the mice receiving therapy-level doses through body weight measurements. In addition, the biodistribution of the ELP following i.t. injection was assessed. During the course of radiotherapy, none of the animals in either tumor model exhibited significant clinical signs of toxicity The treatment of FaDu xenografts with 2000 µCi [¹³¹I]ELP₁₂₀-Tyr₇, for instance, showed a maximum body weight loss of 4% (Figure 4A); generally, a 15% reduction in body weight is accepted before the mouse must be euthanized for humane reasons. Body weight loss in the mice bearing the PC-3 xenograft, however, was more severe, as the maximum body weight loss reached a plateau of 14% on day 15, gradually improving to normal levels over the following two weeks (Figure 4B). However, this toxicity was not primarily caused by the ELP seed, but rather by the PC-3 tumor. Mice bearing the PC-3 tumor displayed greater than 10% weight loss in the absence of treatment (data not shown). The biodistribution results further demonstrate that after i.t administration of this radiotherapeutic depot, clearance of radioiodine from normal tissues was rapid, with the result that by 24 h, activity levels were less than 1% ID/g in all tissues studied including the thyroid (Fig 4C). Autoradiography of the tumor following [¹³¹I]ELP₁₂₀-Tyr₇ infusion illustrate that the ¹³¹I activity was concentrated primarily at the administration site located in the center of the tumor (Figure 4E), and its distribution pattern matched the area of tumor necrosis seen in the histological tumor section (Figure 4D). These results demonstrate that the [¹³¹I]ELP₁₂₀-Tyr₇ radioactivity was mainly restricted to the tumor, and indicate that the systematic optimization of this local drug delivery depot significantly improved antitumor efficacy and lowered systemic toxicity.

Figure 4. Safety evaluation and localization of the ELP depot in mice.

Mean body weight following i.t. administration of unlabeled ELP or [¹³¹I]ELP₁₂₀-Tyr₇ in nude mice bearing (A) FaDu xenografts and (B) PC-3 xenografts, Data represent the mean and SEM, n = 7–12. (C) The biodistribution of radioactivity in mice bearing FaDu tumor xenografts treated with an ELP depot. Data represent the mean and SEM, n = 5–6. (D) H&E stained photomicrograph of the tumor injection site 48 hrs after i.t. administration of [¹³¹I]ELP₁₂₀-Tyr₇ at 50 × magnification, and (E) autoradiography of the corresponding tumor injection site of (D)

Discussion

In this study, we developed an alternative modality to conventional brachytherapy consisting of a biodegradable, genetically engineered polypeptide conjugated to a therapeutic radioisotope that spontaneously transitions from an easily infusible liquid into a seed-like depot upon injection into a solid tumor. We observed that the integrity and efficacy of this depot correlated with the number of tyrosines in its C-terminus tail, overall MW and solution concentration, and demonstrated that these spontaneously assembled seeds are effective against two human tumor xenografts in athymic mice. We hypothesized that the physicochemical and structural properties of the polypeptide seeds could be altered to maximize depot retention and efficacy.

To investigate this hypothesis, we recombinantly synthesized a series of polypeptides, prepared ELP-radionuclide conjugates, and investigated the effect of polypeptide MW, amino acid composition, and injection concentration —the three main parameters controlling their phase transition behavior— on the stability of the depot. Interestingly, we found that each parameter played an important role. Increasing the injection concentration and increasing the number of tyrosines in the sequence increased tumor retention half-life. Increasing the molecular weight also increased tumor retention, though only up to a point; doubling the molecular weight from 50 to 100 kDa did not result in a further increase in seed stability. Importantly, the increased retention time directly correlated with the ΔT_t, defined as the difference between the transition temperature of the ELP under these specific conditions and the intratumoral temperature. We hypothesized that by using these parameters to drive the absolute transition temperature lower —thereby increasing the ΔT_t— we effectively increase seed stability by altering two molecular parameters: viscosity and internal microstructure. We have two lines of evidence suggesting that viscosity and the internal microstructure are important. The first is that the most successful depot, ELP₁₂₀-Tyr₇, also exhibits the highest viscosity (Figure 1B). The markedly different viscoelastic behavior of this depot compared with the two other ELPs studied herein is probably related to the fact that its internal microstructure is quite different from the other two ELP depots. The second piece of evidence was revealed by DLS: ELP₁₂₀-Tyr₇ forms spherical micelles at temperatures below its coacervation temperature, unlike ELP₁₂₀-Tyr₁ and ELP₁₂₀-Tyr₄ (Figure 1C). Thus, the microstructure of this ELP seed may consist of aggregated nanoscale micelles, while the other ELP seeds simply consist of entangled polymer chains with little internal structure. This difference in microstructure, we believe, is largely responsible for the stark difference in the temporal retention of the seed within the tumor, and hence its efficacy in both confining the radioactivity to the tumor space —thereby limiting systemic toxicity— and enhancing efficacy via its ability to irradiate the tumor from the inside out.

Given that this spontaneously formed seed was stable for prolonged periods of time, we hypothesized that it could prove an effective therapeutic for the treatment of tumors in regions accessible by injection such as the head, neck, breast, brain, and prostate. This seed-like depot displayed significant antitumor efficacy against two human tumor xenografts as seen by the following results: 1) 100% of the tumors responded to treatment, resulting in a 67% cure rate; 2) 8/12 mice bearing FaDu tumors and 13/13 mice bearing PC-3 tumors survived until the end of the 60-d experiment. The tumor response was separated into two phases, which is typical of clinical brachytherapy. In the first stage (3–5 days post treatment), the tumor mass was large and rigid, which was likely due to edema of the tumor tissue. In the second stage, approximately 3–7 days post-infusion, the tumor size decreased and the mass became more pliant. Histological analysis indicated that this effect was likely due to necrosis because significant tumor tissue destruction in and around the injection site was observed 72 h post-infusion of the radiolabeled ELP.

We also examined the toxicity of this system in preliminary fashion by monitoring body weight of animals receiving [¹³¹I]ELP. Similar to clinical brachytherapy, this modality was well tolerated by mice implanted with two different human tumor xenografts, suggesting that the therapeutic load was predominantly contained within the tumor. The biodistribution results demonstrated that only a limited amount of the radioactivity accumulated in important organs (Fig 4C). Noteworthy, there were three forms of radioactivity distributed in the tumor and in normal organs, including free ¹³¹I, free [¹³¹I]ELP and aggregated [¹³¹I]ELP, and each one of these is expected to have a different elimination pathway. Free ¹³¹I is rapidly cleared once released from the ELP into circulation; free [¹³¹I]ELP may be cleared off at a slower rate as it diffuses away from the ELP depot and enters systemic circulation. The mechanism of clearance of these two forms may vary depending on its degradation rate and MW, but our biodistribution data (Fig. 4C) suggests a gradual excretion process through the kidneys, as indicated by a higher percentage of the injected dose in the kidneys compared with other organs at each time point. The toxicity contributed by this clearance process will only be minor, however, because we observed that the exposure rapidly dropped below 0.5 % ID/g for all organs after the first 24 hours after administration (Fig 4C). This also allows us to assume in our subsequent analyses that >99% of the detected tumor radioactivity originates from the aggregated [¹³¹I]ELP that constitutes the major radioactive source in our injectable system. These encouraging results suggest that this polypeptide seed may by worthy of further evaluation for several potential applications: 1) to treat patients who are at an early stage (e.g. breast or prostate) without requiring them to undergo radical surgery; 2) to make unresectable tumors operable by debulking their mass (e.g. tumors near major arteries or organs); and 3) to improve the quality of life for patients with advanced tumors (e.g. lung cancers that obstruct the bronchi).

Despite the efficacy of this procedure, we are currently investigating methods to further improve its therapeutic potential. For example, this treatment strategy relies on a uniform coverage of the entire tumor because the penetration depth of the selected radionuclide (¹³¹I) is only 1 mm. While this short penetration depth can prevent harmful radiation from affecting nearby healthy tissues, it can also leave live tumor cells at the tumor boundaries (20). To improve the uniformity of the coverage it may be possible to inject multiple seeds (similar to clinical brachytherapy) to compensate for large or multifocal tumors or modulate the distance the infusion solution travels before transitioning into a stable seed-like depot. Furthermore, providing a second infusion following the initial response may prevent the tumor reemergence that we observed in a few isolated animals.

In conclusion, we have developed a genetically engineered, thermally responsive peptide polymer that self-assembles into a biodegradable seed upon intratumoral injection. By modulating the physicochemical and structural characteristics of the polypeptide, we were able to significantly increase the retention time and the antitumor efficacy of the self-assembled seeds. The intratumoral administration of the seed-like depot led to a 100% response rate and a >67% cure rate in two xenograft models after 60 days, and showed only minimal accumulation (<1 %ID/g) in vital organs. This novel self-assembled seeds are now being studied in orthotopic animal models to assess its feasibility as an alternative to brachytherapy seeds.