Liposomal phenylephrine nanoparticles enhance the anti-tumor activity of intratumoral chemotherapy in a preclinical model of melanoma

Abstract

Intratumoral injection of anti-cancer agents has limited efficacy and is not routinely used for most cancers. In this study, we aimed to improve the efficacy of intratumoral chemotherapy using a novel approach comprised of peri-tumoral injection of sustained release liposomal nanoparticles containing phenylephrine, which is a potent vasoconstrictor. Using a preclinical model of melanoma, we have previously shown that systemically administered (intravenous) phenylephrine could transiently shunt blood flow to tumor at the time of drug delivery, which in turn improved anti-tumor responses. This approach was called the dynamic control of tumor-associated vessels. Herein, we used liposomal phenylephrine nanoparticles as a “local” dynamic control strategy for B16 melanoma. Local dynamic control was shown to increase the retention and exposure time of tumors to intratumorally injected chemotherapy (melphalan). C57BL/6 mice bearing B16 tumors were treated with intratumoral melphalan and peri-tumoral injection of sustained release liposomal phenylephrine nanoparticles (i.e., the local dynamic control protocol). These mice had statistically significantly improved anti-tumor responses compared to melphalan alone (p = 0.0011), whereby 58.3% obtained long-term complete clinical response. Our novel approach of local dynamic control demonstrated significantly enhanced anti-tumor efficacy and is the subject of future clinical trials being designed by our group.

Graphical Abstract.

Introduction

For a wide range of malignancies, most anti-neoplastic agents are administered via a systemic (intravenous) route. In contrast, there is limited use of intratumoral agents mainly because the drug distribution is not predictable or uniform via intratumoral injection. Needle puncture into tumors can be imprecise and somewhat subjective based on the healthcare provider who is administering the injection. In addition, the injection of the anti-tumor treatment may leak out of the tumors and thus decrease the effectiveness of the therapy. Thus, drug administration into the bloodstream and distribution through the tumor-associated vessels is often the most common standard of care approach for drug delivery.

Systemic drug delivery must reach tumor targets through the tumor-associated vasculature, which is widely considered as a common denominator for limiting treatment success. There is considerable tumor vessel heterogeneity in structure and function that limits drug responses and, in turn, correlates with treatment outcomes.^1–5 The most successful chemotherapeutic agents and targeted inhibitors require adequate distribution to the tumor via the circulation, or else they are rendered ineffective.^6–10 Similarly, cell-based onco-immunotherapeutics, such as adoptive T-cell transfer and CAR T cells, are also dependent upon accessing the tumor via the tumor-associated vasculature, and better outcomes have been directly correlated to increased immune cell infiltration into the tumor.^11–13

Our group has previously shown that alteration of tumor blood flow at the time of systemic chemotherapy delivery can enhance chemotherapeutic effects.¹⁴ This was achieved through properly timed administration of a systemic fluid bolus and phenylephrine through an approach we have termed the “dynamic control” of tumor associated vessels. We tested dynamic control in our mouse model of isolated limb perfusion for melanoma, which has been shown to replicate treatment for in-transit melanoma.^15,16 Our approach of dynamic control resulted in shunting of blood flow to the tumor-associated vessels, which are known to have several aberrancies that limit blood flow and therefore hinder the blood-based delivery of systemically administered anti-tumor agents. These limitations include elevated intratumoral interstitial pressure and non-functional tumor vessels, which do not support adequate blood flow.^17–20 In our preclinical animal model, the ability of dynamic control to shunt blood flow to tumor targets at the time of drug delivery resulted in increased drug penetration and significantly improved the chemotherapeutic effects on tumor growth, complete clinical response, and melanoma-specific survival.¹⁴

Intratumoral phenylephrine and other vasoactive agents combined with intratumoral chemotherapy have been used for the treatment of cutaneous head and neck squamous cell carcinoma (SCC) with significant local tumor response.^21–24 Aberrancies in the tumor vasculature of head and neck SCC tumors are similar to those in melanoma, which contribute to limitations in drug delivery through the systemic circulation.^25,26 Taken together, there is an established, albeit limited, role of intralesional therapies with phenylephrine or other vasoactive agents for cutaneous tumors. However, the use of sustained release formulations of vasoactive agents (including phenylephrine) has not yet been tested. In addition, the peri-tumoral (as opposed to intratumoral) use of vasoactive agents has likewise not been tested.

In this present study, we sought to build upon our approach to systemic dynamic control by developing a type of “local” dynamic control. To accomplish this, we constructed sustained release liposomal nanoparticles containing phenylephrine, a potent vascular vasoconstrictor (α agonist). During systemic dynamic control, the effects of intravascularly delivered phenylephrine were short-lived (approximately 5 minutes) due to the brief intrinsic half-life of this drug.²⁷ For local dynamic control, we postulated that a slow release liposomal formulation of phenylephrine could generate sustained local vasoconstriction when injected around the peri-tumoral tissues. When combined with intratumoral chemotherapeutic injections, we hypothesized that our new local dynamic control approach would “entrap” the intratumorally injected agents within the tumor through sustained vasoconstriction of peri-tumoral associated blood vessels. This, in turn, would result in increased dwell or exposure time of the drug to the tumor by decreasing the influx of tumor blood flow and thus decreasing the efflux of the intratumorally injected drug as well.

By addressing the known limitations of intratumoral injections (poor drug distribution or leaking of drug out of the tumor), our goal was to improve anti-tumor effects through our peri-tumoral local dynamic control strategy comprised of sustained release liposomal phenylephrine nanoparticles. To this end, our primary hypothesis was that the combination of intratumoral chemotherapy with peri-tumoral injection of vasoactive liposomal phenylephrine nanoparticles would improve intratumoral chemotherapy retention and efficacy, leading to enhanced anti-tumor responses.

Materials and Methods

Animals

Male and female C57BL/6 mice (6–8 weeks old) and female BALB/c (also 6–8 weeks old) mice were purchased from the Jackson Laboratory (Bar Harbor, ME). Prior to experimentation, mice were allowed to acclimate to the animal facility for at least 1 week. Animal care and use were in accordance with Mayo Clinic institutional guidelines and approved under Mayo Clinic IACUC protocol A00003049-17: Effects of tumor vessel control on systemic anti-cancer therapies.

Tumor cell lines

B16 (subclone F10) melanoma cells and 4T1 triple negative breast cancer cells were obtained from ATCC and authenticated. Cells were cultured in RPMI 1,640 (Roswell Park Memorial Institute Medium) supplemented with 10% FCS (fetal calf serum), 2 mM l-glutamine, 100 U/ml penicillin, 50 μg/ml streptomycin, and 50 μM β-mercaptoethanol (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA).

Reagents

FCS was obtained from Gemini Bioproducts (Woodland, CA). RPMI 1640, PBS, penicillin-streptomycin, and L-glutamine were obtained from Life Technologies Inc. (Grand Island, NY). Melphalan and phenylephrine were purchased from Sigma-Aldrich and stored at 4 °C. Melphalan was dissolved in phosphate buffered saline (PBS) containing 1% dimethyl sulfoxide (DMSO) and 0.1% hydrochloric acid (HCl). Phenylephrine was dissolved in PBS. After the liposomal nanoparticles were generated (as explained in the next Methods subsections), these were also diluted in PBS prior to in vivo injection.

Preparation of blank (empty) liposomes and liposomal phenylephrine nanoparticles

Liposomes were formulated using a modified ethanol injection method. First, a mixture of the necessary phospholipids and cholesterol [DOPC or 1,2-Dioleoyl-sn-glycero-3-phosphocholine, 3.67 mg/mL; cholesterol, 0.483 mg/mL; and DSPE-PEG-Ome or 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-maleimide(polyethylene glycol), 0.27 mg/mL] was taken in an ethanolic solution and heated in a water bath at 65°C for 5 minutes. This mixture was slowly injected into pre-heated milli-Q water under magnetic stirring. This resulted in the spontaneous formation of liposomes as the ethanolic lipid solution encountered the aqueous phase. The mixture was stirred for 15 minutes at room temperature. The ethanolic portion in the water was removed by rotary evaporation under reduced pressure, and the volume was topped up with milli-Q water. This completed the process for blank or empty liposomal nanoparticles.

Liposomal phenylephrine nanoparticles were generated using the same method but by adding 1000 ug/mL of the hydrophilic drug phenylephrine to the aqueous phase. The lipophilic phospholipids and cholesterol were added to the ethanolic solution, heated in a 65°C water bath for 5 minutes, and slowly injected into pre-heated milli-Q water under magnetic stirring. The ethanolic portion in the aqueous phase was removed by rotary evaporation under reduced pressure, and the volume was topped up with milli-Q water. Any unentrapped phenylephrine was removed using an Amicon Ultra centrifugal filter with a cut-off size of 3 kD. The resulting liposomal phenylephrine concentrates were collected, the volume was topped up with milli-Q water, and the liposomal phenylephrine nanoparticles were stored at 4°C.

Physical Characterization and Drug Loading of Liposomes

The mean sizes and zeta potentials of empty and liposomal phenylephrine nanoparticles were measured using Malvern Zetasizer Nano ZS at 25°C. The samples were diluted with water for in vitro analysis, which was performed in triplicate. For studying the stability of the empty and phenylephrine-loaded liposomes, liposomes were incubated in autoclaved Milli-Q water, and size measurements were performed to assess nanoparticle consistency for up to 100 days. The entrapment efficiency of phenylephrine was quantified by using UV-visible spectroscopy at 275 nm. Liposome-entrapment efficiency was measured by determining the amount of unentrapped phenylephrine. Briefly, the encapsulated phenylephrine amount was calculated by subtracting the amount of unentrapped phenylephrine from the total drug amount. The phenylephrine-entrapment efficiency was expressed as the percentage of the entrapped amount to the total amount. The phenylephrine-loading efficiency was calculated as the percentage of the encapsulated amount to the total lipid amount.

Transmission electron microscopy (TEM) analysis

The size and morphological features of liposomal phenylephrine nanoparticles were analyzed by transmission electron microscopy (TEM). Briefly, 5 μl of liposomes was placed on a carbon-coated copper grid (glow-discharged for 45 seconds using a TolarnoHavoc Evaporator) for 10 minutes. The excess sample was blotted with Whatman filter paper. The dried coated grids were vacuum-dried, and the electron micrographs of the liposomal nanoparticles were recorded with a FEI Tecnai 12 TEM electron microscope.

In vitro liposomal phenylephrine release studies at various pH levels

In vitro drug release profiles of liposomal phenylephrine nanoparticles at pH’s of 7.4, 5.4 and 4.4 were measured by the direct dispersion method. These pH levels were selected to replicate normal physiologic pH levels (7.4) and more acidic pH levels known to exist within tumors.

Briefly, liposomal formulations containing a known quantity of phenylephrine were subdivided into multiple equal-volume parts in different Eppendorf tubes. Each part was diluted with buffers of different pHs, and the mixed solutions were incubated in a water bath shaker at 37°C. At 0, 30, 60, 120, 240, 360, 480, and 600-minutes time intervals, tubes were taken out and centrifuged at 2000 rpm for 3 minutes by using Amicon filter tubes (Merck Millipore) with 3 kDa molecular weight cut-off, and the released phenylephrine was quantified by using UV-Visible spectroscopy.

The absorbance was measured at 275 nm using the Jasco V-650 UV-visible spectrophotometer. A standard graph was plotted using varying concentrations of phenylephrine and used as a reference. The phenylephrine release efficiency was calculated using the following formula: % of phenylephrine release efficiency = (concentration of released phenylephrine divided by the total amount of entrapped phenylephrine in the liposomes) ×100.

Fluorescence imaging

Fluorescein (AK-FLOR) was obtained from Sigma-Aldrich and diluted 1:50 with PBS for intratumoral injections. Fluorescein-injected mice were anaesthetized via nose cone using 2.5% isoflurane gas for approximately 10–15 minutes and imaged using an IVIS Spectrum (Perkin Elmer) at excitation and emission wavelengths of 460 nm and 515 nm. Images were obtained at the following time points after fluorescein injection: 2, 24, 48, and 72 hours. Levels of fluorescein were presented as radiant efficiency. Mice were euthanized after the last time point.

Mass spectrometry

The LC-MS system consisted of a Waters Acquity H class ultra performance liquid chromatography (UPLC) system, containing a quaternary solvent manager and sample manager-FTN coupled to a Xevo TQ-S mass spectrometer (Waters, Milford, MA) equipped with an electrospray ionization (ESI) source. Data were acquired and analyzed by Waters MassLynx v4.1 software.

The liquid chromatographic separation of melphalan and melphalan-d8 were accomplished using a Poroshell 120 EC-C18 analytical column (2.1×100 mm, 2.7 μm, InfinityLab) protected with Poroshell 120 EC-C18 precolumn (2.1×5 mm, 2.7 μm) at 40°C, eluted with a gradient mobile phase composed of water containing 0.1% formic acid (designated solvent A) and acetonitrile containing 0.1% formic acid (designated solvent B) with a constant flow rate of 0.3 mL/min and a total run time of 9 min. The elution was initiated at 20% solvent B, followed by increasing solvent B from 20% to 70% over 4 minutes, where it was held for 0.5 minutes and returned to 20% B over 0.5 minutes and at the initial conditions for 3 minutes before the next injection. Autosampler temperature was 4°C, and sample injection volume was 1 μl.

The mass spectrometer was operated in the positive ESI mode using capillary voltage 2.9 kV, source temperature 150°C, desolvation temperature 300°C, cone gas flow 150 L/hr, and desolvation gas flow 800 L/hr, using multiple reaction monitoring (MRM) with a dwell time of 0.044 sec. The cone voltages and collision energies were determined by MassLynx-Intellistart, v4.1, software. The MRM precursor and product ions were monitored at m/z 630.42 > 398.17 for melphalan and 854.43 > 105.09 for melphalan-d8.

The primary stock solutions of upamostat (1 mg/ml in MeOH), UK-1 (1 mg/ml in MeOH), opaganib (1 mg/ml in MeOH), and taxol (IS) (1 mg/ml in MeOH) were prepared in 4 ml amber silanized glass vials and stored at −20°C. Working standards were prepared by dilution of the stock solution with 1:1 MeOH:H2O in 1.7 ml microcentrifuge tubes and stored at −20°C.

Weighed, frozen mouse tumor samples were pulverized using the CP02 cryoPREP automated dry pulverizer (110V). Frozen tissue samples were weighed, transferred to homogenization tubes, and homogenized in 3 parts (weight to volume) chilled phosphate buffered saline (PBS) using a bead mill tissue homogenizer (settings: speed- 5; time- 10 seconds; cycles- 3; dwell- 3). The tumor homogenate samples were extracted as follows: 50 μL homogenate was added to 100 μL methanol containing IS (75 ng/ml) in a 1.5 mL microcentrifuge tube. Following vortex mixing (40 sec), the aqueous/organic supernatant was separated by centrifugation (settings: 15,000 rpm; 15 min; 4°C). The supernatant was transferred to glass autosampler vial inserts for LC-MS/MS analysis.

Assessment of tumor responses in vivo

Three perpendicular axes of the tumors were measured approximately every other day using external digital calipers (Control Co.). Tumor volume was calculated using the formula ½ × length × width × height. Mice that died or were euthanized due to morbidity, tumor ulceration, or tumor reaching the size endpoint (2,000 mm³) were classified as events.

Statistical Analysis

Intratumoral drug concentrations and tumor volumes were assessed by a two-tailed Student t test. Tumor growth was assessed by ANOVA (GraphPad Prism, Version 6 software). Error bars represent standard errors of the mean unless otherwise noted. Standard Kaplan–Meier methods were used to perform the time to event analysis. Values of p<0.05 were considered significant.

Results

Construction and characteristics of liposomal phenylephrine nanoparticles

For the physical and chemical characterization of our liposomal nanoparticles, we analyzed both the empty/blank (LNP) and phenylephrine-loaded liposomal formulations (LNP-Phen). We evaluated the physicochemical properties of the liposomes, including hydrodynamic diameter, surface potential, liposomal shape, and size uniformity by dynamic light scattering (DLS) and TEM studies, respectively, along with the stability of the liposomal formulations. The average hydrodynamic diameters of empty and phenylephrine-loaded liposomes were approximately 50 and 70 nm, respectively (Figure 1A). The surface potentials of empty and phenylephrine-loaded liposomes were both negative, specifically −5 ± 2 and −4 ± 1, respectively. The TEM analysis showed that individual particles had circular shapes and were uniformly distributed (Figure 1B).

Figure 1.

Physicochemical characterizations of phenylephrine loaded liposomal formulations. (A) Hydrodynamic diameters of blank (empty) liposomal nanoparticles (LNP) and liposomal phenylephrine nanoparticles (LNP-Phen). (B) Transmission electronic image (TEM) of liposomal phenylephrine nanoparticles. (C) The entrapment efficiency of phenylephrine was quantified by using UV-visible spectroscopy at 275 nm, which showed approximately 30% liposomal encapsulation of phenylephrine within the liposomal nanoparticles. (D) Stability of liposomal phenylephrine nanoparticles at 4°C over different time points. (E) In vitro phenylephrine release profile at different pH levels, ranging from 4.4 to 7.4. The more acidic pH levels were tested to replicate the tumor microenvironment, where pH levels have been shown to be more acidic than normal tissues.

The amount of phenylephrine entrapped within the liposomal nanoparticles was estimated before the in vivo animal studies. Initially, the amount of phenylephrine was set at 1000 ug/ml. Our process to construct liposomal phenylephrine nanoparticles resulted in a respective percentage of phenylephrine entrapped in the liposome of 280 ug/mL (or 28%), as shown in Figure 1C and Table 1.

Table 1. Lipid compositions and drug loading efficiencies of the liposomal nanoparticle formulations:

The blank/empty liposomal nanoparticles (LNP) and liposomal phenylephrine nanoparticles (LNP-Phen) were comprised of DOPC (1,2-Dioleoyl-sn- glycero-3-phosphocholine), cholesterol, and DSPE-PEG-Ome (1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-maleimide polyethylene glycol). Liposomal phenylephrine nanoparticles contained an average of 0.28 mg/ml phenylephrine, which resulted from an average encapsulation efficiency of 28%.

Liposome	DOPC (mg/ml)	Cholesterol (mg/ml)	DSPE-PEG-Ome (mg/ml)	Loading amount of Phenylephrine (mg/ml)	Encapsulated amount of Phenylephrine (mg/ml)	Encapsulation Efficiency (%)
LNP	3.67	0.48	0.27	-------------------	-------------------	-------------------
LNP-Phen	3.98	0.48	0.27	1.00	0.28 ± 0.09	28 ± 1

The stability of the empty and phenylephrine-loaded liposomes was evaluated by DLS up to 100 days in deionized water (Figure 1D). This analysis showed that the formulations were very stable under the specified conditions. In addition, we tested the stability of the formulations in other media, including PBS, fetal bovine serum-free medium, 10% fetal bovine serum-containing medium (which approximates mouse blood serum), and 55% serum-containing medium (which approximates human blood serum). Table 2 summarizes these results, which showed that both formulations were stable among the different media.

Table 2. Hydrodynamic diameter (nm) of the liposomal nanoparticle formulations in different mediums:

We tested the stability of the two liposomal nanoparticle formulations (LNP – blank/empty liposomal nanoparticles; LNP-Phen – liposomal phenylephrine nanoparticles) in various mediums, including PBS, fetal bovine serum-free medium, 10% fetal bovine serum-containing medium (which resembles mouse blood serum), and 55% serum-containing medium (which resembles human blood serum). Diameter calculations were performed at time 0 and 24 hours after formulation. Formulations were stored at 4°C. We observed a slight increase in the average (avg) hydrodynamic diameters of the particles in PBS, serum-free medium, and 10% serum-containing medium. However, in 55% fetal bovine serum-containing medium, we detected several diameter peaks due to the interference of the protein content in high proportion (55%) of FBS.

Medium	Sterile water		Phosphate-buffered saline (PBS)		Fetal bovine serum-free medium		10% fetal bovine serum medium		55% fetal bovine serum medium
Time	0 hours	24 hours	0 hours	24 hours	0 hours	24 hours	0 hours	24 hours	0 hours	24 hours
LNP (nm) Avg ± std	68 ± 12	70 ± 8	92 ± 12	92 ± 8	95 ± 3	96 ± 5	92 ± 14	95 ± 8	Multiple peaks	Multiple peaks
LNP-Phen (nm) Avg ± std	78 ± 9	79 ± 3	95 ± 13	96 ± 6	94 ± 9	95 ± 7	92 ± 12	96 ± 11	Multiple peaks	Multiple peaks

We also determined the phenylephrine release profile from the liposomal phenylephrine nanoparticles at different pH levels, including 4.4, 5.4, and 7.4, using the direct dispersion method at varying time intervals. Our observations validated that there was a sustained release of phenylephrine from the liposomal nanoparticles, whereby the physiologic pH (i.e., 7.4) showed the slowest rate of release (Figure 1E), which would be expected to reflect in vivo conditions within the peri-tumoral normal tissues. However, the tumor microenvironment has been shown to have lower pH levels than normal (non-tumor) tissue.^28–30 Therefore, we included lower pH levels in anticipation that some of the peri-tumoral injection of the liposomal phenylephrine nanoparticles would reach tumor tissue upon or after injection. We observed more rapid rates of phenylephrine release at lower pH levels. Overall, after the entire analysis period (10 hours), all of the phenylephrine was released from the liposomal nanoparticles regardless of the pH level.

Liposomes increase retention of intratumoral injections consisting of fluorescent dye

To determine whether local dynamic control could increase the retention of intratumorally injected agents, we used two different mouse cancer models. These included B16 melanoma and 4T1 triple negative breast cancer. Tumors were injected orthotopically into either the dorsal skin fold (B16 melanoma) or into one of the inferior mammary fat pads (4T1 breast cancer). A total of 1 × 10⁵ B16 or 4T1 cells were injected in 100 μl of PBS with a 23-gauge needle. When the tumors reached approximately 10 mm³, we injected 50 μl of diluted (1:50) fluorescein intratumorally with or without peri-tumoral injection of liposomal phenylephrine nanoparticles. As a liposome control, nanoparticles that did not contain phenylephrine (i.e., empty or blank liposomes) were also used. For the peri-tumoral injections, 4–6 injections were performed to cover as much tissue surrounding the tumor volume as possible. Injections created a raised wheal surrounding the tumors, and these were also performed superficially and deep to the tumor within the subcutaneous tissues. Each liposomal injection consisted of 100 μl of nanoparticles (blank control or phenylephrine-containing). Peri-tumoral liposomes were injected prior to the intratumoral fluorescein. After 5 minutes of the peri-tumoral liposomal injection to allow more uniform distribution of the nanoparticles around the tumor, the intratumoral fluorescein was administered.

Fluorescence imaging was performed to record radiant efficiency following the intratumoral fluorescein injection at the following time points: 0 hours, 2 hours, 24 hours, 48 hours, and 72 hours. Experiments were performed with 2 mice per group and repeated in triplicate. Figure 2 shows the radiant efficiency levels of intratumoral fluorescein for B16 melanoma (Figure 2A) and 4T1 breast cancer (Figure 2B). At each time point, the intratumoral fluorescence was significantly greater for tumors that received peri-tumoral injection of liposomal phenylephrine nanoparticles compared to tumors that received no liposomes or blank liposomes. There was an expected decrease in fluorescence intensity with time, but the relative amount of fluorescence was still highest at each time point for tumors that were treated with peri-tumoral liposomal phenylephrine nanoparticles. Figure 2C shows an example of the radiant efficiency among 4T1 tumors within an intramammary fat pad in BALB/c mice. Of note, there was minimal detectable fluorescence signal outside of the site of injection, suggesting that little to no fluorescein was released into the systemic circulation with the intratumoral injection. Fluorescence signals extended beyond the tumor borders, suggesting that there was spread of the dye past the tumor itself, which led to larger areas of dye infiltration.

Figure 2.

(A) C57BL/6 mice bearing B16 melanoma tumors within the dorsal skin fold were injected with intratumoral fluorescein (50 μl of a 1:50 dilution) and peri-tumoral injection of either (1) no liposomal nanoparticles, (2) blank/empty liposomal nanoparticles, or (3) liposomal phenylephrine-containing nanoparticles (10 μg per 100 μl per injection). Peri-tumoral injections (4–6 in total) were given along all 6 borders of the palpable tumors (when tumors reached approximately 10 mm³), creating wheals of injected nanoparticles. Fluorescence was measured at 2, 24, 48, and 72 hours after treatment. The highest radiant efficiency at each time point was recorded for tumors treated with liposomal phenylephrine-containing nanoparticles, demonstrating that these sustained release nanoparticles functioned as a “local” dynamic control mechanism to increase prolonged retention of intratumoral fluorescein. (B) Similar findings to (A) were detected in a 4T1 triple negative breast cancer model using BALB/c mice. Breast tumors growing orthotopically within the mammary fat pad had the highest retained fluorescence at all time points when treated with liposomal phenylephrine nanoparticles. (C) Examples of BALB/c mice treated with liposomal phenylephrine nanoparticles or blank liposomes, demonstrating radiant efficiency intensity over the 4 time points of the experiment.

Liposomal nanoparticles containing phenylephrine increase the retention of chemotherapy (melphalan) in primary melanoma

Following intratumoral injection of fluorescein with or without peri-tumoral injection of liposomal phenylephrine nanoparticles, we investigated whether our approach of local dynamic control could increase retention of intratumorally injected chemotherapy. For this, we utilized the B16 melanoma model. Again, 1 × 10⁵ B16 melanoma cells were inoculated within the dorsal skin fold of C57BL/6 mice. When tumors reached 5 mm³, 20 μg of melphalan (in 100 μl solvent as described in the Methods) was injected intratumorally. Four mice were used per group (i.e., no liposomes, blank liposomes, and phenylephrine liposomes), and tumors were resected at the same 4 time points used in the fluorescence imaging experiments (i.e., 2 hours, 24 hours, 48 hours, and 72 hours). Mice were euthanized at the time of surgical resection. The levels of melphalan within the tumors were quantified using mass spectrometry. This experiment was completed in triplicate.

Figure 3A shows the concentrations of melphalan (standardized in ng per mg of resected tumor tissue) for the different groups at the 4 time points. Means and standard deviations represent 3 repeated experiments. Similar to the radiant efficiency measurements, the injection of peri-tumoral liposomal phenylephrine nanoparticles significantly increased the intratumoral concentration of melphalan at all 4 time points. Again, there was an overall expected decrease in intratumoral melphalan concentrations over time, but the level of intratumoral melphalan when given with peri-tumoral liposomal phenylephrine nanoparticles still significantly persisted compared to the controls.

Figure 3.

(A) Mass spectrometry analysis of intratumoral melphalan among B16 tumors treated with peri-tumoral injection of either (1) no liposomal nanoparticles, (2) “blank” liposomal nanoparticles, or (3) liposomal phenylephrine nanoparticles (10 μg per 100 μl per injection). Melphalan was injected into tumors at 20 μg in 100 μl. Tumors were resected at the same predefined endpoints (2, 24, 48, and 72 hours). Measurement of melphalan concentrations (standardized as ng per g of tumor tissue) was performed, which showed that the highest intratumoral melphalan concentrations were achieved with peri-tumoral injection of liposomal phenylephrine nanoparticles. (B) Mass spectrometry analysis was also performed on systemic tissues and peripheral blood at the time of necropsy. Very minimal levels of melphalan were detected systemically at only the 2 hour time point. Melphalan was not detected within these tissues at any other time point, and no levels were detected within the peripheral blood at any time point. These data suggest that there was minimal systemic spread or “leaking” of melphalan from the intratumoral injections.

To determine whether the intratumoral injections resulted in any detectable systemic levels of melphalan, we performed necropsy of all mice after the tumors were resected for mass spectrometry. Full necropsy did not yield any identification of gross melanoma (pigmented) metastases. Tissue samples were taken surgically from the spleen, kidney, liver, and lung. Peripheral blood was also collected. All samples were then analyzed via mass spectrometry for melphalan concentrations. Only at the 2 hour time point were there detectable, but very miniscule, levels of melphalan (as shown in Figure 3B). At all of the remaining time points, melphalan was undetectable among these tissues. At all time points, there was no detectable levels of melphalan within the peripheral blood samples. Together with the fluorescence imaging data, mass spectrometry analysis and necropsy also supported that intratumoral injection with melphalan did not produce any significant systemic spread of the intratumoral chemotherapy or of the B16 melanoma itself.

Peri-tumoral injection of phenylephrine (aqueous formulation) has modest effect on anti-tumor responses with intratumoral melphalan

In this set of experiments, we tested whether peri-tumoral injection of phenylephrine in aqueous solution (standard non-liposomal formulation) could have an effect on intratumoral injection of melphalan (20 μg per 100 μl) with respect to anti-tumor responses. For subcuticular tumor inoculation, 1 × 10⁵ B16 cells were injected in 0.1 ml of PBS. Treatments were initiated when tumors reached approximately 5 mm³, which occurred approximately 10–14 days post-inoculation. As controls, mice receiving only peri-tumoral injection of phenylephrine or only intratumoral melphalan were included. Intratumoral melphalan and peri-tumoral phenylephrine were administered every 3 days. For the peri-tumoral injections, 4–6 injections of phenylephrine (10 μg per 100 μl per injection) were performed to cover as much tissue surrounding the tumor volume as possible and to create multiple wheals of fluid underneath the peri-tumoral tissue. Tumor measurements were taken every 5–6 days. Equal numbers of male and female mice were used for each experiment. Experiments were completed in triplicate to assess for reproducibility.

As shown in Figure 4, peri-tumoral injection of non-liposomal phenylephrine combined with intratumoral melphalan did not have any statistically significant response on tumor growth (Figure 4A) or survival (Figure 4B) when compared to intratumoral melphalan alone. Intratumoral melphalan alone or in combination with peri-tumoral phenylephrine significantly slowed tumor growth compared to negative and phenylephrine controls, as to be expected. Only 1 mouse (12.5%) showed complete response in the group that received combined intratumoral melphalan and peri-tumoral phenylephrine. There were no complete responses in the melphalan alone group. No significant differences were identified between male and female mice.

Figure 4.

Peri-tumoral injection with phenylephrine dissolved in aqueous solution (standard formulation) did not have any effect when combined with intratumoral melphalan on anti-tumor responses, including tumor growth (A) or survival (B), when compared to intratumoral melphalan alone. In this experiment, C57BL/6 mice bearing orthotopic B16 tumors were treated with peri-tumoral aqueous phenylephrine. Controls included phenylephrine and melphalan only injections. While melphalan with or without peri-tumoral phenylephrine generated statistically significant improvements in tumor growth as expected, standard phenylephrine did not significantly enhance anti-tumor effects when compared to melphalan alone. (C) Example of a B16 tumor ulceration as confirmed on IHC on a C57BL/6 mouse, which developed during treatment. The combination treatment with intratumoral melphalan and peri-tumoral phenylephrine increased local toxicity in the form of tumor ulceration. Ulceration was a common adverse event, with 75 to 87.5% of mice demonstrating tumor ulceration that necessitated local wound care.

Of note, mice treated with peri-tumoral phenylephrine with or without intratumoral melphalan showed increased ulceration of tumors compared to negative and melphalan only controls. In groups receiving phenylephrine, 75 to 87.5% of mice (6–7 out of 8 per group) demonstrated tumor ulceration during the course of treatment, typically starting after 3 to 4 treatments. Ulcers were treated with topical lubricants and antibiotics (bacitracin) per our IACUC protocol exemptions and also led to earlier euthanasia (predefined endpoint) for affected mice. This suggested that this combination approach with intratumoral melphalan and peri-tumoral phenylephrine increased local toxicity.

At the end of these experiments, necropsy was performed on all mice to evaluate for the presence of metastases. No metastases were identified in any mice.

Peri-tumoral injection of liposomal phenylephrine nanoparticles significantly improves anti-tumor responses with intratumoral melphalan

To test our primary hypothesis that peri-tumoral injection of liposomal phenylephrine nanoparticles could function as a local dynamic control mechanism, phenylephrine nanoparticles were combined with intratumoral melphalan. Similar to the previous experiment, 4–6 peri-tumoral injections of liposomal phenylephrine nanoparticles (10 μg per 100 μl per injection) were given with intratumoral melphalan (20 μg per 100 μl) every 3 days. Treatments were initiated when tumors grew to approximately 5 mm³. Injection of nanoparticles alone were included as a control. Equal numbers of male and female mice were used for each experiment, and experiments were completed in triplicate.

Unlike the aqueous (standard) formulation of phenylephrine, the liposomal phenylephrine nanoparticles had a significant effect on tumor growth (Figure 5A) and survival (Figure 5B) with a complete clinical response achieved in 58.3% (7/12) of mice treated with the combination of intratumoral melphalan and peri-tumoral liposomal phenylephrine nanoparticles. Mice that obtained complete tumor response, of which 3 were female and 4 were male, did not have tumor volumes that exceeded 25 mm³ during the course of the experiment. This suggested that tumor growth kinetics beyond 25 mm³ was sufficient to overcome the anti-tumor effects of our local dynamic control strategy. Treatment ended 60 days after tumor inoculation, and these mice with complete response showed long-term survival (over 90 days) and were then humanely euthanized.

Figure 5.

Peri-tumoral injection with liposomal phenylephrine nanoparticles enhanced the effects of intratumoral injection of melphalan, as demonstrated by decreased tumor growth (A) and improved survival (B) compared to intratumoral melphalan alone. Mice treated with the combination of liposomal phenylephrine nanoparticles and intratumoral melphalan had a 58.3% (7/12 mice) complete response rate when compared to controls (including melphalan only controls). Tumor ulceration still developed for mice treated with liposomal phenylephrine, though to a lesser extent with the standard formulation (33.3 to 50% of mice per group that received liposomal phenylephrine nanoparticles).

Mice treated with liposomal phenylephrine nanoparticles also showed evidence of tumor ulceration. However, this occurred to a lesser extent with 33.3 to 50% of mice (4–6 out of 12 per group) having demonstrated tumor ulceration. For mice that achieved tumor regression and complete response as well as tumor ulceration, ulcers healed about 15 days after termination of treatment with new skin and scar tissue formation. At the end of these experiments, necropsy was performed on all mice, and no metastases were identified.

Discussion

Although the intratumoral route is not the mainstay for the delivery of cancer therapies, chemotherapy still has an important role in the treatment of cutaneous malignancies, including melanoma. Melphalan, a DNA intercalating agent, has long been shown to have anti-tumor effects in vitro and in vivo for both preclinical animal models and in humans.^14,31–34 Its main clinical use has been for the treatment of melanoma in-transit disease during regional isolated limb infusion or perfusion.³⁵ The effects of melphalan during regional therapies can be quite significant, and it is often used for those patients who progress on immunotherapy or who otherwise are not candidates for immune checkpoint blockade (e.g., patients with organ transplants). Thus, in our study, we developed a novel approach to utilize the anti-tumor effects of melphalan through a new local, peri-tumoral strategy consisting of liposomal phenylephrine nanoparticles.

The concept of intratumoral injection of chemotherapy by itself is not a novel one. Innovative agents, including small molecules and other nanoparticles, have been tested via the intratumoral route of administration and have shown significant promise in preclinical models of melanoma.^36,37 Investigators have combined intratumoral melphalan with other strategies, including sonoporation or electroporation, in order to increase the concentration of melphalan within melanoma tumors.^38,39 These techniques have been employed in other cancer models that utilize intratumoral injections as well.^40,41 Other novel agents have been tested via the intratumoral route of administration, which have also shown promise in preclinical models of melanoma.^42–44 Furthermore, studies in humans have shown the benefit of intratumoral chemotherapy for various human malignancies. For example, intravitreal injection of melphalan has been successfully used for the treatment of intraocular retinoblastoma.^45–47 Other intratumoral injections are available for melanoma, including Talimogene Laherparepvec (T-VEC). T-VEC is a herpes simplex virus oncolytic therapy that selectively replicates within melanoma cells.^48,49 It produces granulocyte macrophage colony-stimulating factor (GM-CSF), which in turn enhances systemic anti-tumor immune responses to melanoma-associated neoantigen targets. T-VEC can be injected into cutaneous melanoma tumors, but it can also be used for regional and distant metastases as long as these can be accessed via image-guided percutaneous injections.⁵⁰ Collectively in these studies, intratumoral injection was found to be a safe and effective delivery mechanism without increased incidence of systemic metastases generation.

Local injection of phenylephrine has also previously been investigated and found to enhance anti-cancer treatments.^51,52 For example, one preclinical study of intratumoral epinephrine and fluorouracil in a mouse model of sarcoma demonstrated superior responses with this combination compared to fluorouracil alone.⁵³ In addition, intratumoral phenylephrine and other vasoactive agents combined with chemotherapy have been used for the treatment of cutaneous head and neck squamous cell carcinoma (SCC), which also demonstrated significant local tumor responses.^22–24 Phase 3 clinical trials have validated the use of injectable vasoactive gels combined with intratumoral chemotherapy for SCC as well.^21,54 Similar to our preclinical animal results, patients who achieved partial or complete pathologic responses in these clinical trials were more likely to have smaller tumors, suggesting that the growth kinetics of the tumor may impact the success of intratumoral therapy. In these clinical studies, it is likely that the vasoactive agents functioned in a similar manner to our approach of local dynamic control by entrapping the drug within the intratumoral tissues. This effect potentiated by vasoactive agents increased retention or exposure time, which subsequently allowed for a more prolonged period for the drug to take effect within its tumor targets.

Building upon prior studies that utilized intratumoral injections, we developed an innovative strategy to enhance intratumoral chemotherapy using sustained release peri-tumoral injection of liposomal phenylephrine nanoparticles. Similar to intra- and peri-tumoral injections of vasoactive drugs in either liquid or gel forms, our constructed vasoactive nanoparticles likely produced a prolonged vasoconstriction that enabled the increased entrapment of the injected chemotherapy within the tumor (as shown in Figures 2 and 3). Our approach of local dynamic control was similar to our previous findings of systemic dynamic control in that both manipulated the tumor-associated blood vessels to optimize drug delivery.¹⁴ Tumor-associated vessels are known to be very heterogeneous and are characterized by several aberrancies that impact blood flow into the tumor. These include haphazard structural arrangements, nonfunctional vessels that do not support blood flow, increased interstitial pressure that limits blood flow, and increased permeability that may alter the distribution of drugs within the tumor.^6,55–57 Addressing the tumor-associated vasculature can be accomplished with a variety of approaches, including anti-angiogenic agents that target the tumor neovasculature (like bevacizumab) or novel antibody-drug conjugates that target tumor vessel receptors and facilitate increased drug delivery.^58,59 In our study, we have focused on the physical parameters of the blood flow through tumor-associated vessels via vasoactive manipulation with our sustained release phenylephrine nanoparticles. These physical parameters, including blood vessel diameter and blood flow velocity, are critically important for drug delivery, which in turn influences anti-tumor responses as shown by our group and others.^11,14,60

Notably, our local dynamic approach did result in increased toxicity in the form of tumor ulceration. Given the vasoactive effects of phenylephrine, this adverse effect was not entirely unexpected. Some of the side effects of prolonged vasoconstriction are ischemia and hypoxia, which in turn lead to necrosis of the affected organs. Necrosis or ulceration can affect the skin, which may result from subcutaneous or intravenous injections.^61–63 Tumor ulceration was noted to be higher with the aqueous formulation of phenylephrine as compared to the liposomal nanoparticle formulation. This was likely because the aqueous solution of phenylephrine exerted all of its effects immediately as opposed to the slow release nanoparticle formulation, which would be expected to exert its effects in a more gradual time-release manner. The dose of phenylephrine used in our experiments (10 μg per 100 μl injection) was selected from our previous animal study using the same dose as a continuous infusion.¹⁴ However, the additive dose of the phenylephrine (4–6 injections or up to 40–60 μg per treatment) over the course of administering injections every 3 days was likely too toxic for the animals in our melanoma model. Indeed, as we formulate early clinical trials that investigate the translational potential of local dynamic control in human melanoma, a safe dose of liposomal phenylephrine nanoparticles will need to be established.

There are many future directions and applications of local dynamic control for the treatment of melanoma and potentially other malignancies. While we utilized a novel sustained release formulation of phenylephrine, we used standard melphalan (aqueous solution) for the intratumoral chemotherapeutic injections. Others have shown that a sustained release liposomal formulation of melphalan enhanced the anti-tumor effects in preclinical animal models.^64–66 Our group and others have also published on the beneficial effects of other sustained release liposomal chemotherapies.^67,68 Furthermore, liposomal doxorubicin is an effective and approved drug for different human cancers, including breast and ovarian cancer.⁶⁹ Thus, combining sustained release liposomal formulations of intratumoral chemotherapy may likely have an additional effect on anti-tumor responses when used in combination with our local dynamic control approach. In addition, while we have focused our investigation of local dynamic control on melanoma, there are other cutaneous or superficially located tumors that may benefit from our approach. We showed that our local dynamic control strategy also worked to increase fluorescein dye retention within 4T1 breast cancer tumors grown orthotopically within the mammary fat pad of mice (Figure 2B and 2C). Future experiments with breast-specific chemotherapies (such as standard or liposomal formulations of doxorubicin) would likely reproduce our findings in melanoma. It is reasonable that local dynamic control could be applied to any accessible tumor target that can be injected by gross visualization, palpation, or assisted with image guidance (e.g., ultrasound or CT scan). As we work to translate local dynamic control to human disease, we are developing a clinical trial for primary resectable melanoma that utilizes local dynamic control in a “neoadjuvant” approach that will determine responses prior to surgical resection.

We recognize that there are limitations to our local dynamic control strategy. As discussed, there was significant toxicity in the form of tumor ulceration with our approach, which would need to be addressed with optimal dosing prior to clinical use. In our proposed clinical trial, patients would still undergo resection of the primary melanoma (to 1–2 cm depending on the Breslow depth), and so the area of ulceration or necrosis (should it develop) would also be resected as part of the surgery. Also as discussed, our local dynamic control strategy seemed to have the most beneficial anti-tumor effects on smaller B16 tumors (approximately 25 mm³ or less). But if B16 tumors grew larger, then the growth kinetics appeared to outpace our treatment. With larger primary tumors or tumors that are located in anatomical deep spaces, it would be more challenging to completely surround the tumor with liposomal phenylephrine nanoparticles, which may limit treatment effect. In addition, there will likely be a baseline variability with the quality of the intra- and peri-tumoral injections, which would be provider-dependent. However, this is a known limitation of all injectable anti-cancer therapies (such as T-VEC).

In conclusion, we have developed a novel approach to local dynamic control that utilized liposomal phenylephrine nanoparticles as a method to increase the retention time of intratumorally injected chemotherapy. Local dynamic control enhanced anti-tumor responses with statistically significant and quite dramatic improvements in complete response and prolonged survival. While there was an increased rate of tumor ulceration as an adverse event, this complication can be addressed in future studies that better optimize the dose of phenylephrine within the injectable sustained release nanoparticles. Importantly, our strategy of local dynamic control offers an effective means to improve upon chemotherapeutics and may translate into alternative clinical treatment approaches for patients who progress on other therapies.

Abbreviations

ANOVA

analysis of variance

cm

centimeter

DLS

dynamic light scattering

DOPC

1,2-Dioleoyl-sn-glycero-3-phosphocholine

DSPE-PEG-Ome

1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[maleimide(polyethylene glycol)

FCS

fetal calf serum

GM-CSF

granulocyte macrophage colony-stimulating factor

kDa

kilodaltons

KM

Kaplan-Meier

mm

millimeter

ng

nanogram

nm

nanometer

PBS

phosphate buffered saline

rpm

rotations per minute

RPMI

Roswell Park Memorial Institute Medium

SCC

squamous cell carcinoma

TEM

transmission electron microscopy

T-VEC

Talimogene laherparepvec

μg

microgram

μl

microliter

μm

micrometer

UV

ultraviolet