Application of an assay Cascade methodology for a deep preclinical characterization of polymeric nanoparticles as a treatment for gliomas

Abstract Glioblastoma multiforme (GBM) is the most devastating primary brain tumor due to its infiltrating and diffuse growth characteristics, a situation compounded by the lack of effective treatments. Currently, many efforts are being devoted to find novel formulations to treat this disease, specifically in the nanomedicine field. However, due to the lack of comprehensive characterization that leads to insufficient data on reproducibility, only a reduced number of nanomedicines have reached clinical phases. In this context, the aim of the present study was to use a cascade of assays that evaluate from physical-chemical and structural properties to biological characteristics, both in vitro and in vivo, and also to check the performance of nanoparticles for glioma therapy. An amphiphilic block copolymer, composed of polyester and poly(ethylene glycol; PEG) blocks, has been synthesized. Using a mixture of this copolymer and a polymer containing an active targeting moiety to the Blood Brain Barrier (BBB; Seq12 peptide), biocompatible and biodegradable polymeric nanoparticles have been prepared and extensively characterized. In vitro studies demonstrated that nanoparticles are safe for normal cells but cytotoxic for cancer cells. In vivo studies in mice demonstrated the ability of the Seq12 peptide to cross the BBB. Finally, in vivo efficacy studies using a human tumor model in SCID mice resulted in a significant 50% life-span increase, as compared with non-treated animals. Altogether, this assay cascade provided extensive pre-clinical characterization of our polymeric nanoparticles, now ready for clinical evaluation.


Figures:
: Schematic representation of the in vitro BBB models. BBMVECs are cultured on the collagen coated insert surface and astrocytes are grown below, on the bottom of the multi-well plate. Figure S2: 1 H-NMR of an example batch of the P co-polymer. Figure S3: GPC of three big batches of the polymer P, at a) initial preparation time (t=0) and b) after 4 months stored at -20ºC. Figure S4: 1 H-NMR of an example batch of the polymer 2P, at a) initial preparation time (t=0) and b) after 6 weeks stored at -20ºC. Figure S5: A -Brain uptake and B -Kin parameter of free PTX and PTX-NPs, functionalized with Seq12 peptide NPs(Seq12) and without any functionalization NPs(NF). **p<0.01. Table S1: Physicochemical characterization of three independent batches of polymer P. Table S2: Physicochemical characterization of three independent batches of polymer 2P. Table S3: Physicochemical characterization of three independent small batches of nanoparticles. Table S4: Physicochemical characterization of three independent big batches of nanoparticles. Table S5: Nanoparticle induced hemolysis (%). n=6replicates/sample. Note that PTX concentrations are much higher than in in vitro experiments because in hemolysis experiments, the expected injected concentration must be tested, and it is always higher in vivo than in vitro.

Tables:
Explanation S1: Characterization of synthesized polymers -Gel permeation chromatography (GPC) Gel permeation chromatography (GPC) was used to determine the molecular weight of the synthesized polymers. Around 1.5 mg of the polymer and polystyrene standards (MW = 2500; 5000; 9000; 17500; 30000 Da; calibration curve) were disolved in 1 mL tetrahydrofurane. 20 µL of each sample were analyzed on a GPC KF-603 column (3 m, 6.0x150) from SHODEX with a flow rate of 0.5 mL/min of tetrahydrofurane. The

(Equation 5)
-1 H -NMR Proton nuclear magnetic resonance ( 1 H -NMR) was performed to determine the ratio of PEG in the polymers P, P* and 2P and nanoparticles. In brief, around 10 mg of polymer P, P* and nanoparticles were dissolved in 0.8 mL deuterated chloroform, while polymer 2P was dissolved in dimethyl sulfoxide. 1 H -NMR spectra was recorded with 16 scans on a Varian 400 MHz. Data were processed using the MestReNova software. The ratio between the PEG and the polyester was determined for the polymers P, 2P or nanoparticles by integrating the signal of PEG (3.64 ppm) and that of polyester (4.06 ppm), as described by Equation 1. To obtain the percentage of peptide in the polymer 2P, the signals of PEG (3.64 ppm) and tyrosine (6.60 ppm) were integrated. The number of protons of the signal of PEG was 256 and for the signal of tyrosine was 2. In order to assess the barrier integrity, our BBB model was characterized in terms of trans-endothelial electrical resistance (TEER) and permeability (Pe) of the water-soluble molecular tracer called Lucifer Yellow (LY). Moreover, the expression pattern of tight junctions' proteins was studied by immunofluorescence techniques.  LY is a fluorescent dye that cannot be taken up by EC in vivo, neither by active nor by facilitated transport; thus, the study of its permeability coefficient (Pe) is used as a marker for layer integrity and BBB quality of the in vitro models, together with the determination of the TEER values. As shown in Figure S3  Tight junctions (TJ) are known to be the main BBB physiological feature. TJs extend circumferentially around the endothelial cells forming a barrier to paracellular passage of small hydrophilic molecules such as sodium, hydrogen, bicarbonate, and also restricting the movements of proteins, particles like viruses and even cells across the BBB. They also play a role in endothelial cells polarization restricting the movement of membrane molecules between the apical and basolateral membrane surfaces.
In order to phenotypically assess our in vitro BBB model, a specific analysis of the expression pattern of tight junctions' proteins, such as ZO-1, Claudin-5 and Occludin, was carried out using immunostainning techniques. A monoculture of BBMVECs was used as a control. As shown in Figure S3.3, the expression pattern of these three proteins was increased when BBMVECs were co-cultured with glial cells. This explanation aims to describe the final point criteria for animal sacrifice. These parameters are included in the protocol number 4565 approved by Direcció General del Medi Natural, Generalitat de Catalunya and have been translated from Spanish.

PARAMETERS OF SUPERVISION IN EXPERIMENTAL ONCOLOGY
Weight loss 0) Normal weight 1) Less than 10% 2) Between 10 and 15% in a few days 3) Consistent or fast, more than 20% maintained for 72 hours.
3) Abdominal distension. Ascitic liquid volume greater than 10% by weight initial body. Breathing difficult (particularly if accompanied by nasal discharge and / or cyanosis). Caquexia.

Alterations in behavior
0) None 1) Incapacity to move normally 2) Inability to reach food / drink. Isolation of the rest of the cage animals 3) Unconscious or comatose. Intent to "hide" on the chip, does not respond to stimuli (dying) Wounds 0) None 1) Scratches

2) Wounds that do not heal
3) Ulcerated wounds that can even be supposed. Ulcerated tumors or necrotic

TOTAL
When the circumstance is that there is more than one parameter with a value of 3 automatically all 3 will pass to 4.
The following monitoring criteria will be taken into account: 1. From 0 to 6 points: Individual assessment of the animal. Possible existence of pain or anguish. The use of painkillers must be considered 2. 6 to 12 points: Existence of pain or anguish. Administration of palliative analgesics.
3. From 12 to 18 points: Compulsory use of painkillers. The sacrifice of the animal must be considered.
4. From 18 to 24 points: Obligated sacrifice of the animal. The end of the procedure must be considered.
Similarly, a final point criteria will be applied at any time during the study if: -The size of the tumor mass, although not reaching a size considered critical (3 in the  assessment table), influences the other bodily functions or causes pain and / or prolonged suffering. -Loss of body weight greater than 20% with respect to basal weight or a control. Explanation S6: Study of the stability of polymer P.

Experimental procedure
The polymer P (batch SAG022-044) was divided in three portions and each one was kept at room temperature, 5 ºC or -21 ºC. A sample was taken at different times and analyzed by GPC to determine the molecular weight and by NMR to calculate the ratio of PEG.
The ratio of PEG in the polymer was calculated from the ratio of the signal of polyester (4.06 ppm) and the signal of PEG (3.64 ppm) in the 1 H-NMR.
The amount of acid of the polymer was determined by titration with KOH. 20 mg nanoparticles with 3wt% of the polymer 2P and 2wt% of paclitaxel were prepared. The size, pdi and zeta potential of the nanoparticles were obtained by Dynamic Light Scattering (DLS). The drug content was determined by UPLC analysis (protocol SAG-WI-035).
The stability of three different batches of polymer P was studied in order to validate the first study with one polymer. The polymer batches SAG022-050, SAG022-068 and SAG022-087 were stored at -20 ºC and samples were taken at different times. The GPC, NMR and acid content was determined as in the previous study. 20mg of nanoparticles with 3wt% of the polymer 2P and 2wt% of paclitaxel were prepared and analyzed by DLS and UPLC. The amount of PEG of the nanoparticles was calculated by the 1 H-NMR of the freeze-dried nanoparticles.

Results
One portion of the polymer SAG022-044 was stored at room temperature and samples were taken and analyzed at different times. The results of the experiments are summarized in Table S6.1.    Table   S6.2. The molecular weight decreased after 3 months but the chromatograms were quite similar for all the samples ( Figure S6.2). The amount of acids increased slightly after 2 months but less than the sample stored at room temperature. The nanoparticles obtained from the polymer showed similar size and zeta potential up to 3 months. In this case, the experiments indicated that polymer P was reasonably stable at 4 ºC for a short time storage.  Table S6.3. Table S6.3: Characterization of the polymer P as a function of time, at -20ºC.
The molecular weight for the polymer stored at -20 ºC was quite stable in time and the chromatograms were very similar ( Figure S6.3). The acid amount was constant all the time and the nanoparticles had quite similar size and zeta potential. This data indicated that polymer P was enough stable at -20 ºC in the period of the study.

Conclusions
The polymer P was kept at -20 ºC and no changes were observed in the capacity to obtain nanoparticles from it at least for 4 months. This is the most appropriate temperature for a long storage of the polymer.
The polymer could be stored at 4 ºC for 2 months. It is important to avoid the storage of the polymer at room temperature, because at this temperature the degradation of the polymer is faster. Explanation S7: Study of the stability of polymer 2P.

Results
Three batches of polymer 2P SAG023-033, SAG023-070 and SAG023-084 were stored at -20 ºC and samples were taken and analyzed by NMR at different times. The results of the experiments are summarized in Table S7.1. A small reduction in the amount of peptide for the polymer SAG023-070 after 1 month was observed. On the other hand, the ratio of polyester had not a considerable variation.
This data could indicate a small degradation of the polymer SAG023-070 after 1 month.
But the other polymers SAG023-084 and SAG023-091 seem to be more stable.  The ratio of polyester and PEG in the polymer had no variation with time and the percentage of peptide was quite similar in the period of study. We can say that the solution of polymer 2P in DMSO was stable at -20 ºC after 6 weeks at the concentration of study.

Conclusions
The polymer 2P can be stored in solid state or solution in dimethyl sulfoxide (12 mg/mL) at -20 ºC at least for 6 weeks.

Explanation S8: PTX release studies
Previous work on paclitaxel (PTX) release profile characterization suggested that the widely used dialysis membrane method was not appropriate due to poor solubility of PTX in aqueous media, and crossing problems through the dialysis membrane when a hydrotropic agent is used for its enhancement. Thus, an alternative method was applied based on nanoparticles filtration and centrifugation. Experiments were performed with the presence of human serum albumin (HSA) and without it, in order to evaluate differences between drug release profiles without and with the presence of proteins.
Briefly, 10 mL of nanoparticles (6 mg/mL, drug content: 1.45%) were placed in a Falcon tube in PBS media or HSA (6 mg/mL in PBS) solution. Both nanoparticles suspension were incubated at 37°during 1 hour, 4 hours and 24 hours, and samples were removed at the corresponding time points. All samples were analyzed by UPLC (Waters Acquity UPLC H-Class), by using the method described in Explanation S2.
Recovery of PTX was between 70% and 90%. The released fraction of the drug is represented in Figure S7.1. Drug release profile shows a normal increasing tendency with time. While drug release performed in PBS resulted in more than 25% drug release at 24 hours, this value is decreased to less than 15% when test is performed in HSA, suggesting than HSA takes an active role in drug release event. Explanation S9: Nanoparticles long-term stability study

Schematic working conditions of the nanoparticles freezing
In Figure S9.1, the conditions at which nanoparticles have been submitted to find out the best storing condition are schematized. Figure S9.1: Schematic representation of the conditions at which nanoparticles have been submitted to find out the best storing condition.

1) NP stability to freeze-drying
Many studies have described freeze-drying as an appropriate methodology to store nanoparticles in solid state. For this reason, in the present work, nanoparticles were also freeze-dried after a freezing step at -20ºC and -80ºC, and then, their sizes were measured. Results are presented in Table S9.1. Table S9.1: Hydrodynamic diameter (in nm) of nanoparticle dispersions after being freeze-dried (freezing step at -20 or -80ºC for 1, 4 and 24hours and thawing at room temperature) As Table S9.1 shows, independently of the freezing time and freezing temperature, nanoparticles were not stable at after freeze-drying. Their sizes increased dramatically, which was expected according to their visual appearance (they showed a more turbid and viscous aspect). Sonication for up to 10 minutes after thawing did not improve the results obtained, indicating that these nanoparticles could not undergo a freeze-drying process.

2) NP stability as a function of the temperature a. At 25ºC and 4ºC as a function of dispersant media
Nanoparticles were dispersed either in water or PBS, without and with 10wt% of sucrose. Then, they were stored at 25ºC and 4ºC. At specific time points, their size, PDI and surface charge were measured. In addition, the pH of the media was also followed as a function of time, since it could be an indication of the nanoparticles hydrolysis.
Results are presented in the following Figures, separately to facilitate the discussion.
First, the results of the pH as a function of time are described in Figure S9.2. As expected, the pH of the nanoparticle dispersions depended on their dispersant: when dispersed in water ( Figure S8.2a), nanoparticles suffered an initial pH decrease down to pH around 4 -4.5. However, after around 35 days, the pH suffered another decrease, independently on the storing temperature and on the presence of sucrose. Nevertheless, nanoparticles dispersed in 10wt% sucrose, at 4ºC, showed the highest stability. In contrast, when nanoparticles were dispersed in PBS ( Figure S8.2b), due to the buffer nature of it, the pH of the dispersion was maintained constant at pH around 6.5 -7, for the whole period of time studied.
It is worth noting that further pH decreases, specifically when nanoparticles are dispersed in water are not expected, because nanoparticles are composed of carboxylic acids, which maximum pH decreases down to around 4. water or b) PBS, without and with 10wt% sucrose, at 25ºC or 4ºC.
As Figure S9.3 shows, initially, nanoparticles showed hydrodynamic diameters around 125 nm. All NPs formulations maintained their sizes over time independently on the dispersion media and the storing temperature up to 20 days. Then, a size increase was observed for all nanoparticles, but in a higher extent when they were stored at RT.
Therefore, nanoparticles are stable in terms of size for at least, 20 days and the most stable nanoparticles are those preserved at 4ºC, dispersed in water + 10wt% sucrose (up to 30 days).  Regarding surface charge, as Figure S9.5 shows, after some initial variations, it is more or less constant over time, but it depended on the dispersant media more than on the temperature. When nanoparticles were dispersed in water, they showed surface charges around -5 mV, while the surface charge was around -11 mV when they were  To sum up this section, it can concluded that nanoparticles dispersed in water + 10wt% sucrose have the highest stability. Therefore, this dispersant medium will be selected for further storage experiments. At 4ºC, as expected, nanoparticles were more stable than at room temperature.

b. Stability of frozen NP as a function of freezing/thawing speed and freezing temperature
Freezing / thawing speed is a parameter defined in previous bibliography to influence the stability of nanoparticles once frozen. For this reason, diverse conditions of freezing / thawing were tested. First, the studies were intended to achieve nanoparticles freezing and storage at -20ºC. Nanoparticles were frozen in a fast way (directly frozen to -20ºC) and in a slow way (decreasing the temperature 1ºC/min until -20ºC). Then, they were thawn in a fast way (directly at room temperature, let thawing for 30 min) and in a slow way (30 min at 4ºC followed by 30 more min at room temperature).
After this procedure, size, PDI and surface charge were characterized after 1 day and 7 days of freezing. In addition, the paclitaxel concentration was also quantified. Results are presented in Table S9.2. concentration of paclitaxel (in mg/mL) of nanoparticles after freezing at -20ºC and thawing at room temperature using different methodologies, for 1 and 7 days. Values in red indicate results markedly deviated from those specified in the specifications.
The initial size of these nanoparticles was around 111 nm, with a PDI around 0.115, their surface charge around -4 mV and the paclitaxel concentration around 1 mg/mL. After 1 day of freezing, independently on thawing and freezing method, nanoparticles sizes slightly increased up to around 130 nm. The PDI was maintained in the range to define nanoparticles as monodisperse; the surface charge, except for one value, slightly decreased in absolute value to around -3mV and the concentration of paclitaxel was maintained as the initial one. Therefore, these results seem to indicate that, using the conditions defined in this experiment, nanoparticles were stable, for at least 1 day, at -20ºC.
However, after 7 days frozen, results were very different, independently of the freezing / thawing method used. Nanoparticles sizes increased notably up to diameters around 200 nm, thus producing an increase on the PDI value. Although the surface charge and paclitaxel concentration seemed to be maintained, these nanoparticles were not stable to freezing 7 days.
Since at -20ºC nanoparticles were not stable for long periods of time, their stability at -80ºC and -150ºC was also tested, firstly during one week. The same parameters were tested. Initially, these nanoparticles showed hydrodynamic diameters of around 120nm, PDI typically lower than 0.2, surface charges between -2 and -6 mV and concentrations of paclitaxel around 1mg/mL.
As Table S9.3 shows, nanoparticles sizes, PDI, surface charges and paclitaxel concentration were maintained similar to those found initially, independently on the freezing temperature and the freezing method, for at least, one week. concentration of paclitaxel (in mg/mL) after freezing at -80 and -150ºC and thawing at room temperature for 1, 3 and 7 days. Values in red indicate results markedly deviated from those specified in the specifications.
The results using a fast freezing are very similar between -80ºC and -150ºC, therefore, -80ºC was chosen to perform further studies on nanoparticles freezing. Accordingly, nanoparticles physicochemical properties after freezing at -80ºC, using a fast freezing were determined for a longer storage time. As Table S8.4 shows, all the studied parameters (size, PDI, z potential and PTZ concentration) kept constant as a function of time for, at least, 50 days (Table S9.4). Therefore, these nanoparticles can be stored frozen at -80ºC for nearly 2 months. Table S9.4: Hydrodynamic diameter (in nm), including PDI; surface charge (in mV) and concentration of paclitaxel (in mg/mL) in nanoparticle dispersions after freezing at -80ºC and thawing at room temperature, as a function of time.
In addition, since -80ºC seemed an optimal temperature to freeze nanoparticles, but -20ºC did not, it was tried to perform a first freezing at -80ºC for 1h, 4h and 24h, followed by the nanoparticles freezing at -20ºC. Nanoparticles characterization was performed at various time-points, as specified in Table S9.5. As this Table shows, all parameters kept constant for, at least, 11 days (longer times under study). Therefore, after a short freezing at -80ºC, nanoparticles can be frozen at -80ºC. Table S9.5: Hydrodynamic diameter (in nm), including PDI; surface charge (in mV) and concentration of paclitaxel (in mg/mL) in nanoparticle dispersions after freezing at -80 and -20ºC and thawing at room temperature, as a function of time. Table S9.6 shows the results of an example of the filtration of a nanoparticles big batch through 0.22 µm and 0.45 µm and its post characterization. As this table shows, nanoparticle size, PDI and zeta potential was maintained constant after nanoparticle filtration. Therefore, the filtration did not produced changes in nanoparticles.   As Figure S2 shows, a large peak appears at 3.6 ppm, which is attributed to PEG (methylene groups); as well as a multiplet at 4.1 ppm, attributed to the CH2 groups of Table S5: Nanoparticle induced hemolysis (%). n=6replicates/sample. Note that PTX concentrations are much higher than in in vitro experiments because in hemolysis experiments, the expected injected concentration must be tested, and it is always higher in vivo than in vitro.