We set out to develop a multimodal nanoparticle platform, based on oil-in-water nanoemulsions carrying iron oxide nanocrystals for MRI, as well as Cy5.5 or Cy7 fluorescent dye for NIRF imaging and the hydrophobic glucocorticoid prednisolone acetate valerate (PAV) for therapeutic purposes (). To synthesize the nanemulsions all the components, i.e. soybean oil and lipids, were dissolved in chloroform. When appropriate, oleic acid coated iron oxide nanocrystals, rhodamine lipid, Cy5.5-DSPE or Cy7-PEG-DSPE lipid and/or PAV were included as well. The mixture was added dropwise to a 70 °C Hepes buffer under vigorous stirring to accomplish the immediate evaporation of chloroform and the formation of a crude oil-in-water emulsion. Subsequently the formulations were homogenized and sized by sonication, washed with fresh Hepes buffer and concentrated using Vivaspin filter tubes.
In the different nanoparticle formulations and their composition are shown. The morphology of the nanoparticles, with and without iron oxide nanocrystals, was investigated with TEM (). The hydrodynamic diameter of the different nanoemulsions was measured with dynamic light scattering (DLS). As shown in , the inclusion of additional functionalities and/or materials did not result in a change of diameter for the different nanoparticles (around 50 nm) or polydispersity. This relatively small size facilitates the extravasation of the nanoemulsions from the circulation over the tumor endothelium into the interstitial space, but also accommodates nanoparticle uptake by cells.
The drug inclusion in the final formulations was determined via spectrophotometric measurements. We determined the PAV inclusion efficiency, established for 11 different nanoemulsion formulations, to be 67.99% ± 9.18 of the initial input value.
The therapeutic effect of the different nanoemulsion formulations was investigated in vitro using a bioluminescent viability assay at a dose of 100 μg PAV/ml and a range of 270-380 μg Fe/ml. To study the acute and long term effect on cell viability two incubation schemes were applied: a 6 hours incubation followed by washing and 18 hours of cell growth, and a 24 hours of incubation. Macrophages, endothelial and tumor cells were chosen since they represent important cellular components of a tumor.
The effects were expressed in % of viable cells compared to untreated control cells. After 6 hours of exposure, LS174T tumor cells were the least affected (). A more pronounced effect at the same time point was observed for J774A1 macrophage cells () and HUVEC () after incubation with PAV-nanoemulsions, RGD-PAV-nanoemulsions and PAV+FeO-nanoemulsions. At 24 hours LS174T cells showed a significantly reduced viability after incubation with FeO and PAV+FeO-nanoemulsions only (). Both J774A1 and HUVEC were very responsive to the treatment and displayed a very low viability after 24 hours of incubations with FeO- and PAV+FeO-nanoemulsions. A loss of viability was also observed in J774A1 and HUVEC that were incubated with PAV-nanoemulsions and RGD-PAV-nanoemulsions ().
To visualize nanoemulsion uptake by the different cell types in vitro
MRI of cell pellets () and fluorescence imaging of well plates were performed (Figure S1
). We observed the different cell types to avidly engulf the different nanoemulsion formulation, albeit at different levels depending on the cell type and nanemulsion formulation used. Quenching effects, in both the MRI and the fluorescence imaging measurements, complicate the interpretation of the data and therefore cannot serve as a quantitative outcome.24
To enable a fully quantitative evaluation of nanoemulsion uptake of the different formulations by the different cell types T1 values (ms) were measured by Minispec benchtop NMR on HCl digested cell pellets. The relaxation rate R1 (1/T1, s−1
) directly correlates with the iron concentration. As expected, the highest R1 values were found for J774A1 macrophages incubated with FeO- and RGD-FeO-nanoemulsions; 0.823 s−1
and 0.811 s−1
, respectively. The R1 of HUVEC incubated with RGD-FeO-nanoemulsions was 0.784 s−1
and 0.577 s−1
when incubated with FeO-nanoemulsions, revealing the elevated uptake of RGD-functionalized nanoemulsions by endothelial cells, known to over-express the ανβ3
For the LS174T tumor cells, also known to express ανβ3
we also observed different R1 values, 0.618 s−1
and 0.695 s−1
when incubated with FeO- and RGD-FeO-nanoemulsions, respectively. These data confirm that our nanoparticle platform functionalized with RGD peptides specifically interact with ανβ3
integrin expressing cell types.
In addition to the investigation of therapeutic response, which we will discuss in detail in the forthcoming sections, we performed MRI on the 4 mice selected (median sized tumors) from groups that were injected with CTRL-nanoemulsions, PAV-nanoemulsions, PAV+FeO-nanoemulsions and FeO-nanoemulsions, using a 3 T clinical scanner. Representative T2*-weighted images of the aforementioned groups are presented in . Tumors of animals injected with PAV-nanoemulsions () and CTRL-nanoemulsions () appeared bright as compared to surrounding muscle tissue. On the other hand the tumors of FeO+PAV- and FeO-nanoemulsion injected mice, shown in , respectively, appeared hypointense as compared to the tumors of mice injected with nanoemulsions that did not contain FeO, indicative of FeO accumulation in these tumors. We quantitatively evaluated the mean T2*-values of the tumors of the different groups. To that end we generated T2*-maps from the T2*-weighted images with different echo times (Figure S2
). displays the T2*-map values and shows these in tumors of animals injected with nanoemulsions containing FeO to be 50% reduced compared to mice injected with nanoemulsions that did not contain FeO.
The inclusion of Cy7 NIR dye coupled lipid in the nanoemulsion corona enabled us to perform in vivo
NIRF imaging to further corroborate the delivery and localization of the nanopartilces at the level of the entire animal. From preliminary ex vivo
NIRF imaging data we found the nanoparticles to accumulate in the liver, kidney and tumors (Figure S3
), with no significant differences between untargeted- and RGD-targeted-nanoemulsion injected mice.
In the full study, four mice injected with PAV-nanoemulsions labeled with Cy7-PEG-DSPE in the lipid corona (Cy7-labeled nanoemulsion) were imaged 48 hours after the injection and prior to sacrifice. Four mice injected with PAV-nanoemulsions that did not contain the Cy7-PEG-DSPE lipid served as controls. Preferential nanoemulsion accumulation in tumors was appreciated (). The mean photon count tumor/skin ratios obtained for Cy7-labeled nanoemulsion injected mice was 11.29 ± 4.96, compared to 0.83 ± 0.013 for mice injected with unlabeled nanoemulsions ().
In vivo and ex vivo fluorescence imaging
Confocal laser scanning microscopy was performed on tumor sections to visualize nanoemulsion localization in the tumor tissue as well as to investigate differences in distribution between the untargeted and the αvβ3-specific RGD-functionalized nanoemulsions. We observed the latter targeted nanoemulsions to be primarily co-localized with the vasculature (), while the untargeted nanoemulsions were found extravasated from the blood vessels in the tumor space ().
For the treatment studies 56 mice (7 groups of 8 mice) were included. One group served as saline treated control, one group was injected with free PAV, while the other 5 groups were treated with different nanoemulsion formulations. Treatment was initiated when tumors were palpable and the tumor volumes were measured daily using a digital caliper throughout the treatment period. In typical tumors at day 8 are displayed of mice treated with PAV-nanoemulsions, RGD-PAV-nanoemulsions, PAV+FeO-nanoemulsions and CTRL-nanoemulsions. The tumor growth in mice injected with PAV-nanoemulsions, RGD-PAV-nanoemulsions and PAV+FeO-nanoemulsions was significantly inhibited compared to the mice injected with saline, CTRL-nanoemulsions, FeO-nanoemulsions and free PAV at a dose of 10 mg PAV/kg (). At day 8 the mean tumor volume was 243.2 ± 73.5 mm3
for PAV-nanoemulsion treatment, 213.6 ± 49.2 mm3
for RGD-PAV-nanoemulsion treatment, and 232.1 ± 46.5 mm3
for PAV+FeO-nanoemulsion treatment (). At day 8, saline, free PAV, CTRL-nanoemulsion, FeO-nanoemulsion treated animals displayed significantly larger tumor volumes of 471.7 ± 79.18 mm3
, 459 ± 77.7 mm3
, 513 ± 111.45 mm3
, 398.8 ± 59.13 mm3
, respectively (). To corroborate the tumor volumes measured by the digital caliper we determined the tumor weight after the sacrifice () and found a similar pattern. The correlation coefficient was established to be 0.779 (Figure S4
). The data presented demonstrate that our nanoemulsions can be used as an effective drug delivery system for tumor therapy. As shown by the tumor growth profiles and tumor weight measurements, all the PAV loaded nanoemulsions induced a significant inhibition of the tumor growth compared to control groups. After three injections over a period of 8 days, the PAV loaded nanoemulsion groups displayed tumor volumes that were at least 50% smaller than all control and free PAV groups at a dose of 10 mg/kg.
Therapeutic effect of nanoemulsions
Prior to the above study we conducted a pilot experiment where animals were also treated at a dose of 20 mg PAV/kg. This resulted in a 77% tumor growth inhibition (Figure S5
) but was associated with severe weight loss, which motivated us to lower the dosing. At a dose of 10 mg/kg we did not find any evidence of serious adverse effects, which is described in more detail below.
The effect of treatment on body weight was established by daily measurements. No relevant weight loss was observed for all groups, with a maximum, but non-significant, reduction of 11.55% in mice injected with PAV–nanoemulsions at day 8 of treatment (Figure S6
). After sacrifice, kidney and liver weights were recorded and normalized to body weights. When we compared the normalized data of the treated groups versus the saline groups marginal variations were noticed in the kidneys weights for all the different groups and a marginal but not significant reduction of the liver weight/body weight ratio was observed in mice injected with CTRL-nanoemulsions, PAV-nanoemulsions and RGD-PAV-nanoemulsions (data not shown). In all the groups treated with nanoemulsions and free PAV, kidney and liver histologic sections showed similar parenchyma morphology to the saline group, demonstrating that PAV and nanoemulsion administrations induced no toxic effects in these organs ().
Histological analysis of kidney and liver tissues
To clarify the therapeutic effect, both angiogenesis and inflammatory infiltration phenomena were studied. Despite the effects of some nanoemulsions on endothelial cell growth in vitro
, staining and subsequent quantification of blood vessels in the tumors did not reveal significant changes between the different treatment modalities (). This was confirmed by analysis of mRNA expression of important growth factors and cell adhesion molecules involved in angiogenesis, which were not significantly altered following treatment (Figure S7A, B
). To evaluate the infiltration of inflammatory cells into the tumors, immunostaining for CD68, a macrophage marker, was performed on tumor sections (). The area was measured and normalized to the total surface area of tumor sections analyzed. As shown in the graph in , the animals injected with saline, control nanoparticles and FeO nanoparticles induced a lower macrophage infiltration compared to the mice injected with nanoparticle formulations containing PAV, although not all the differences were statistically significant.
Angiogenesis and macrophage tumor infiltration evaluation
Most cancer chemotherapies are characterized by limited efficacy and strong side effects caused by the high and frequent dosing. Therefore there is a strong motivation to develop treatment that enable efficient and safer drug delivery to tumors, and consequently reduce dosages and collateral effects. Nanomedical treatments represent an attractive alternative and were pioneered over three decades ago with the introduction of liposomal drug formulations.28
Currently, some of these formulations are applied in clinical practice,29
but the success of liposomes is hampered by a variety of limitations, including the inability to include water insoluble compounds, size them well below 100 nm and relatively low encapsulation efficiencies. The past decade has seen unprecedented growth in the field of nanochemistry, which has resulted in the availability of numerous new nanomaterials to be potentially used for therapeutic and/or diagnostic purposes.
In a previous study we developed and extensively characterized a new nanoparticle platform that was based on a small oil-in-water nanoemulsion,22
which allows the delivery of hydrophobic agents and nanomaterials. In the present study we modified this platform to obtain “theranostic” nanoemulsions to enable imaging guided treatment of experimental cancer. For diagnostic purposes iron oxide nanocrystals were included in the core and Cy7-conjugated lipid in the corona to enable in vivo
MR and NIRF imaging, respectively. At day of sacrifice (day 8), dark areas were visible in tumors on T2*-weighted MR images acquired in mice injected with FeO- and PAV+FeO-nanoemulsions. NIRF images acquired from mice injected with nanoemulsions containing Cy7-conjugated lipids revealed massive nanoemulsion accumulation in the tumors. We attribute the tumor accumulation of untargeted nanoemulsions to the enhanced permeability and retention effect, a phenomenon that occurs in tissues with a leaky vasculature, such as tumors and inflamed tissue.30,31
On the other hand, RGD-peptide functionalized nanoemulsions accumulated in the tumor tissue by active targeting of tumor blood vessels as revealed by confocal microscopy. Moreover, for therapeutic purpose we included a hydrophobic glucocorticoid (PAV) in the nanoemulsion formulation and tested the pharmacological effect in the same animals. We tested a number of glucocorticoids, but PAV was selected based on previous studies,18
its hydrophobicity and its high inclusion efficiency.
To study active targeting, the cyclic RGD pentapeptide was conjugated to malemide functionalized PEG lipids that were included in the lipidic corona. This peptide is known to have a high affinity for ανβ3
which is over-expressed by angiogenically activated endothelium.34
Vascular targeting is attractive because angiogenesis is required for tumor growth35
and the vasculature is readily accessible for targeting and does not require the nanoparticle to extravasate into the tumor interstitium. In a number of reports by us36-38
this targeting strategy has been shown to be very valuable for target-specific imaging with a variety of nanoparticles, including liposomes, quantum dots and microemulsions. In the present study we observed that in mice treated with RGD-PAV-nanoemulsions the tumor growth was inhibited. Ex vivo
confocal imaging performed on tumor slices showed the nanoparticles colocalized with tumor blood vessels.
Expression of growth factors and receptors involved in angiogenesis measured at the mRNA level, as well as the CD31 positive area, did not show any significant difference between the groups. Based on our data, it seems that process of angiogenesis is not involved in the therapeutic effect. Another aspect that has been studied in relation to the anti-tumor mechanism is the macrophage tumor infiltration. Measuring the percentage of the tumor area with CD68 positive cells showed an interesting difference between mice injected with nanoemulsion containing PAV and the control groups. All the groups treated with PAV loaded nanoemulsions showed the same trend: a higher infiltration of cell expressing CD68 compared to saline and control groups. It has been established before that tumor associated macrophages (TAMs) can have a double function. These TAMs can exert angiogenic effects by secreting cytokines such as IL-1, IL-8, TNF-alpha,41-43
but can also release angiostatic factors such as IL-12, MMP-12 and IL-18.44-46
Therefore their presence is often associated with tumor progression, metastasis and poor prognosis; but in some tumor types, including prostate47
and stomach cancer,48
the presence of TAMs appears to be associated with improved prognosis. Some studies49,50
have shown that tumor endothelial cells have a suppressed expression of adhesion molecules reducing the leukocyte-vessel wall interactions. The present study suggests that the different nanoemulsions do not induce differences in macrophage infiltration by regulating the endothelial adhesion molecule make up. While we did not see any differences in angiogenesis, our data suggest that in this particular tumor model the treatment induced an influx of TAMs that resulted in hampered tumor growth. Further studies are needed to unravel the exact mode of action.