1. Introduction
In recent years, cellulose derivatives have gained renewed interest in the field of green chemistry [
1,
2] and among those, carboxymethylcellulose (CMC) has immense applications in the food, pharmaceutical, and cosmetic sectors [
3]. Pharmaceutical grade sodium CMC is water-soluble at neutral pH, it is available in high purity forms and has found different biomedical applications due to its biocompatibility, rheological properties, and low cost. Given its plant origin, CMC is less likely to cause an immune response, which is a fundamental advantage over other natural additives derived from animals, such as collagen. Furthermore, the digestive enzyme cellulose, cellulase, is absent in humans. This allows for a good in vivo stability, compared to other natural polymeric additives susceptible to enzymatic activity [
4]. Thanks to the carboxylic acid functionality, CMC is a polyanion often used as a stabilizer for the synthesis of inorganic nanoparticles [
5] and, in the field of polyelectrolyte complexes applications (PEC) [
6,
7], it is frequently investigated as a key component of three-dimensional hydrogels [
8] and microparticles [
9].
CMC based nanoparticles (NP) are generally investigated as self-assembled structures from CMC derivatives [
10], hybrids or PEC [
11]. Presently, the displacement reaction of sodium counterions by other metallic ions has been shown to form coordination complexes, investigated for nanomedicine applications [
12]. In this paper, we firstly supposed that this behavior could be exploited for the obtainment of CMC-based NP by ionotropic gelation, in the presence of divalent cations such as Ca
2+. Secondly, similar ion coordination could be used for the inclusion of Gallium-68 trivalent ion (
68Ga
3+), being such a positron-emitting radionuclide for positron emission tomography (PET) imaging. Since 2000,
68Ga
3+ GMP-grade
68Ge/
68Ga generators have been commercialized, thus allowing the in-house production of
68Ga-radiolabeled radiopharmaceuticals also in absence of cyclotron facilities.
68Ga
3+ has advantageous decay characteristics (89% β+ yield, 1.9 MeV; half-life of 67.7 min) conferring increasing opportunities for its use in radiopharmacy [
13]. In particular,
68Ga-PET is thought as a possible alternative for some
99mTc-single photon emission computed tomography (SPECT) applications, due to the better spatial resolution, sensitivity, rapidity of execution, and patient compliance of PET over SPECT imaging [
14]. One typical example of such a transition from SPECT- to PET-based radiopharmaceutical is represented by radiolabeled-somatostatin receptor analogs [
15].
In the present paper, we focused on the radiolabeling of white blood cells (WBCs) as a model of transition from
99mTc-based to
68Ga-radiolabeling, being aware that the potential limitation of
68Ga half-life for late-point imaging acquisition.
99mTc-exametazime (
99mTc-HMPAO) WBCs radiolabeling is a well-established technique, widely available in all nuclear medicine departments. The radiolabeling procedure is executed by means of devices specifically developed for this application and by the handling of radiopharmaceuticals kits, to guarantee the effectiveness of the procedure along with patient and operator safety. WBCs radiolabeling is performed to detect and localize infections or inflammations and it is applied for the evaluation of several disorders, including osteomyelitis, infected joint and vascular prosthesis, diabetic foot, lung infections, neurological infections, etc. [
16,
17].
During the last years, nanoparticulate carriers alongside molecular systems, typically bifunctional chelators, have been progressively applied in diagnostic imaging. The main advantage of the use of nanoparticulate carriers regards the possibility of modulating the interaction with the biological system, by tuning the surface and bulk chemical-physical features, and enhancing the efficacy of the transported active, with feasible applications also as theranostic agents [
18,
19]. In particular, surface features play a major role in mediating the interaction with the biological entities, especially regarding protein interaction and relevant cellular internalization [
20,
21]. In the case of polymeric NP, the endosomal escape is commonly related to the pH responsiveness of the nanocarrier [
22]. In the present studies, we aim to investigate pH-sensitive CMC based NP as radiopharmaceuticals-kit for WBC-labeling. The research activity was performed taking into consideration the need for a simple labeling protocol, quickly executable without risk of potential contamination. Two approaches were pursued, the first involving the simultaneous cross-linking of CMC into NP and
68Ga
3+ loading, readily usable for WBC labeling. The second based on preformed NP, which has to be labeled with
68Ga
3+ under mild conditions, either as NP suspension or lyophilized powder, and then incubated with WBC.
CMC-based NP were prepared by ionotropic gelation and scaled-up. pH-responsiveness, as well as serum stability, was investigated. The optimized formulations were tested for the radiolabeling of freshly isolated WBC under hospital clinical protocols.
3. Results and Discussion
In the field of biomedical imaging, positron emission tomography (PET), with its ability to identify and quantify picomolar quantities of radionuclide, has emerged as one of the most powerful diagnostic techniques. Currently, studies concerning polymeric vehicles for
68Ga radiolabeling are relatively few and involve hybrid systems such as coated inorganic nanoparticles, methacrylate polymers or dendrimers (i.e., PAMAM) [
23,
24]. The choice of CMC for this specific
68Ga-NP application stems from the desire to work with a known product already in pharmaceutical use. This choice could facilitate a faster clinical translation of the performed studies. Additionally, despite being known and used for decades as a rheological excipient for numerous pharmaceutical applications, CMC transformation into nanoparticles by ionotropic cross-linking can equally have interesting scenarios and, to our knowledge, has not been extensively explored, yet. In the present research work, it was possible to prepare two CMC-based nanoparticulate systems: In the first case the trivalent ion nature of
68Ga
3+ was used as co-crosslinker for ionotropic gelation, in the second, instead, the labeling was performed of preformed particles.
3.1. Radiolabeling and In-Situ Formation of CMC Based NP: I KIT-Type Approach
In this first approach, the radionuclide eluted in 0.05N HCL solution is added directly to CMC solutions for the formation of NP by ionotropic gelation, in the presence of Ca2+ as a cross-linking agent. Initially, the optimization of the formulation parameters was related to the formation of an opalescent suspension of NP with uniform diameter distribution and an average diameter below 500 nm. Furthermore, it was necessary to obtain NP suspension suitable for direct incubation with living leukocytes: That means sterility, neutral pH, and isotonicity. Additionally, fast incubation and operator-friendly procedures had been considered. The simple handling and mixing of three solutions were envisaged: A solution of CMC, the acidic freshly eluted solution of 68GaCl3, and a solution of CaCl2 as a consolidating agent. During the preliminary formulation studies, a non-radioactive GaCl3 in 0.05 N HCl solution was used.
As reported in
Table 1, several combinations of CMC/CaCl
2wt ratios were assayed, as well as the use of a buffered solution. Initially, it was observed that NP formation was not occurring in the presence of the sole
68GaCl
3 solution (RUN 1 and 2) due to the extremely low concentration of salt (0.003 nM), and that NP formation was sensitive to the solution pH as expected (RUN 10, 11). For the tuning of the final pH value, the combination of PB and NaOH 1 N was examined considering also that the addition of CaCl
2 lowered the pH and that stronger PB was altering the tonicity of the solution as well as leading to the formation of calcium phosphate precipitates. The formulation named RUN 15, resulted perfectly buffered without any additional pH adjustment and leading to a homogeneous diameter distribution with a mean diameter of 200 nm. The ionotropic cross-linking process is a cooperative process [
25] and requires a consolidation time to stabilize the cross-linked macromolecules into NP. Being in the presence of a radionuclide, rapid consolidation time reduces the loss of radioactivity due to the normal radioactivity decay and it is preferable for operator exposure safety. The optimal consolidation time was found to be 10 min from the addition of CaCl
2 solution. Additional time was not significantly modifying, either the diameter distribution, or the scattering intensity (i.e., 5 × 10
−6 counts/s).
The ionotropic cross-linking of CMC into NP was confirmed by ATR/FT-IR spectroscopy (
Figure 1). The comparison of native CMC and NP evidenced the shifting of the O–H band from 3300 nm
−1 to 3359 nm
−1, including an increased intensity and narrowed amplitude. Furthermore, the carboxylate bands at 1589 and 1416 nm
−1 for the CMC spectrum shifted to 1632 and 1426 nm
−1. These differences are assigned to the presence of calcium ions acting as bridges between two carboxylate moieties, thus cross-linking the CMC. Consequently, the free carboxylate moieties are reduced in number, determining also a variation of the intra-/intermolecular hydrogen bridges between carboxyl and backbone hydroxyls.
The selected formulation was then prepared in the presence of freshly eluted
68Ga
3+ and characterized. Filter sterilized solutions of CMC and CaCl
2 were placed in autoclaved glass bottles and sealed, before being submitted to
68Ga
3+ labeling/NP formation. No substantial differences were observed, in terms of size distribution, yield, and final pH, between the suspensions prepared in the simulated medium (RUN 15) and those prepared in the presence of the radioactive (named In-situNP,
Table 2). The specific radioactivity measured on centrifuged NP resulted in 156 mCi/mg, but the LE% was 11.4.
3.2. Radiolabeling of Preformed CMC NP: II KIT-Type
The second modality investigated for the labeling of NP of CMC concerns the use of a preformed NP sample, to be incubated with the freshly eluted radioactive agent. Similarly, the essential requirements were a homogeneous nanometer-scale diameter distribution, physiological pH and isotonicity of the resulting suspension, rapidity, and simplicity of handling for the operator, good specific radioactivity of the resulting NP. Two approaches were investigated: A concentrated, purified NP suspension (NPSusp) and a lyophilized NP powder (NPLyo). In both cases, it was first necessary to scale up the NP preparation protocol up to 30 times the original RUN14 formulation scheme. The obtained NP suspension (RUN16) have an average diameter of 217 nm with a low polydispersity index (PI: 0.333), in agreement with RUN14 results (
Table 1). NP purification was performed by centrifugation and the conditions were optimized in order to maximize the yield, resulting as 31.4 ± 0.6 wt %, but preserving a nanoscaled diameter distribution (
Ø of purified NP: 296.4 ± 6.3 nm, PI 0.323).
The labeling of NPSusp in terms of RADst and LE% (
Figure 2) was investigated at two different incubation times and by varying NP concentration in solution. Lower NP concentrations displayed higher RADst but lower LE%, in agreement with the dynamics of the supposed complexing/Ca
2+-exchange with CMC carboxylate moieties. The variation of the incubation time from 5 min to 15 min did not improve RADst due to concomitant
68Ga
3+ radioactivity decay. A concentration of 5 mg/mL was selected and applied for the subsequent investigations.
Lyophilization conditions for purified NP were also assayed. Several conventional cryoprotective agents were tested, such as PVP, PEG, and trehalose. The lyophilized samples were then re-dispersed in 0.05 N HCl (mimicking the elution buffer of the radioisotope), and analyzed by DLS. Among tested conditions, 5% trehalose was the most effective cryoprotecting agent and allowed for preserving the nanoscaled diameter distribution (
Table 3). Diversely, the other agents caused a considerable increase in the polydispersion and significant variations in the average diameter value. The physiological pH value of the re-dispersed NPLyo was guaranteed by using Tris base buffer (0.1 M). The organic buffer was preferred over the more common phosphate buffer, thanks to its poor contribution to the solution osmolarity [
26] and to avoid the salting-out effect, otherwise induced by PB during the freeze-drying process.
The radiolabeling of preformed CMC-based NP (the II kit type approach) involved the preparation of both NPSusp and NPLyo by using filter-sterilized solutions, and the execution of the protocol under a laminar flow cabinet. Freshly generated
68Ga
3+ solution was applied, and
Table 2 displays the labeling results of NPSusp and NPLyo. Both formulations led to LE% higher than 80% and comparable RADst. The higher RADst values recorded for NPsusp are ascribable to the wider specific surface of the smaller NP of NPSusp (average diameter of 299 vs. 343 nm of NPLyo).
3.3. NP Stability: Effects of pH Acidification and Serum
The response of purified NP to pH variation was assessed by DLS measurements. The suspension remained opalescent, without significant variation in the size distribution until pH 6.5. Lower pHs corresponded to the loss of stability, with a gradual reduction of NP average diameter, combined with a massive increase of polydispersity (
Figure 3). This behavior is indicative of the loss of cross-linking due to the increasing of carboxylic moieties with respect to carboxylate ones, as approaching polymer pKa (CMC pKa: 4.30). Meanwhile, the loss of solubility of CMC at lower pHs was reflected by the formation of micron-sized aggregates (Ø > 4 μm), with sedimentation tendency.
NP stability in serum is a central reason for their application as WBC-labeling agent. Tris-buffered 5 mg/mL NP suspensions (named as Plain) were used and added either with NaCl under physiological osmolarity concentration or trehalose. CaCl
2 was also added to 0.42 mM final concentration, as typically adopted in cell culturing media suited for WBC [
27], such as the Roswell Park Memorial Institute (RPMI)-1640. The assayed samples represent the potential media in which the WBC internalization tests can be carried out, considering freshly isolated WBC from peripheral blood. The presence of calcium allows for the mediation of the endocytosis process, while sodium chloride and trehalose guarantee an isotonic environment and, for the latter, the efficacy of the lyophilization process.
The interaction study with serum was mandatory for the selection of the best incubation conditions, since CMC can directly interact with the residual serum proteins entrapped in wet WBC isolated pellets, forming with those aberrant coagula which leads to procedure failing. NP stability in serum was correlated with the time needed for the formation of macroscopic clots in the dispersing media, containing 5%
v/
v of CPF [
16]. The samples were also characterized in terms of Zeta potential (ζ), measured in the absence of CPF. The recorded data are plotted in
Figure 4 and a strict correlation was observed between NP stability with serum and the corresponding ζ value. All the zeta potential values were found to be negative, due to the carboxylate moieties of CMC. Taking the plain samples as a reference (ζ −17.2 mV), the absolute value always increased with the addition of inorganic salts. Differently, in the presence of trehalose the absolute ζ value decreased, remaining negative, anyway. Trehalose is a non-ionic disaccharide, therefore, it does not act as a counter-ion within the electrical double layer on the NP surface, but its effect is more easily ascribable to non-specific adsorption. Hence, there is a screening action toward the surface charges and a consequent reduction of the absolute ζ value. Although further investigations are necessary, it is possible to state that the extension of stability timing in serum is strictly linked to the surface properties of NP in suspension. The screening effect of more stable NP with high ζ potential values, as well as adsorption of trehalose on the colloid surface, resulted in a delayed coagula formation.
To summarize, the pH sensitivity of the CMC-based NP and the tendency of CMC to interact with serum proteins were confirmed. The optimization of the formulation aspects was necessary to limit the formation of coagula during the subsequent studies of cellular uptake. However, both aspects appear useful and promising for the modulation of the intracellular trafficking of CMC based NP. In the present application, the pH-sensitivity can be useful for the endosomal escape by exploiting the natural pH acidification of endosome maturation, thus leading to a reduced radioactivity efflux for radionuclide exchange through cellular proteins interaction.
3.4. WBC Labeling
The two kit type approaches for WBC radiolabeling were compared through the WBC internalization assessment. The In-situNP formulation was tested, involving the concomitant formation of NP suspension and labeling guided by the ionotropic gelation mechanism. The second approach, concerning the labeling of preformed NP, was investigated for both NPSusp and NPLyo formulations. The WBC internalization tests were carried out by using WBC isolated from donor peripheral blood. The internalization was monitored for incubation periods up to 45 min and was evaluated in terms of radioactivity recovered in WBC (INT%). The permanence of the radioactivity (RES%) in labeled isolated WBC was also measured, suggesting partial radioactivity outflow over time (
Table 4).
Concerning In-situNP, I procedure approach, INT% is low and it was not time-dependent. Indeed, In-situNP presented a large amount of free CMC, compared to that truly converted to NP (33 wt %), which was not significantly internalized by WBC giving no contribution to cell labeling. Additionally, it was not possible to calculate the outflow due to the low residual radioactivity in cell pellets, which was not enough for a reliable reading.
Regarding the preformed NP, II procedure approach, both NPSusp and NPLyo were tested. NPSusp excipients were Tris buffer, CaCl
2, and NaCl, similarly to NPLyo, except for the switch of NaCl with trehalose. INT% was measured at 15, 30, and 45 min but NPSup failed visual inspection after 20 min of incubation, due to the formation of macroscopic aggregates and indicating non-sufficient stability in the presence of the cellular pellet. Considering the specific hospital guidelines [
16], the formation of precipitates compromises the successful use of labeled-WBC suspension, being unsuited for patient re-infusion.
Conversely, NPLyo passed both visual inspection and trypan blue exclusion test [
16] with less than 1% of cell damage during the entire study (efflux period included). The internalization data display increasing radioactivity over time, suggesting for time-dependent cellular uptake. The samples incubated for 45 min were submitted to the efflux study, monitoring the radioactivity of WBC after 15 and 45 min. A reduction of RES% over time is observed, following radioactive outflow. The residual percentage radioactivity at 45 min was equal to 52%. The assessment of longer periods was not performed because of the rapid radioactivity decay of
68Ga (t
1/2: 68 min) and considering that at least 90 min (sum of 45 min incubation time and 45 min for efflux evaluation) passed since
68Ga had been eluted from the generator.
Indeed, the surface characteristics of the nanoparticle (charge density, presence of hydrophobic domains, stealth molecules or directional molecules) and the cellular environment are very important and correlated factors [
28]. Typically, cationic nanoparticles carry a better interaction with the negatively charged cell membrane. On the other hand, the opsonization process generally mediates the uptake of negatively charged NP, with a significant contribution depending on the type of adsorbed protein. It is interesting to note that even the protocol applied to cell isolation can affect the mechanisms of NP uptake. Baumann et al. (2013) reported that, in the case of negatively charged NP, the uptake from leukocytes can strictly depend on the anticoagulant type and that heparin can significantly increase the internalization compared to those positively charged NP. In the present work, an interaction between NP and serum was observed, leading to the formation of macroscopic coagula. However, the working conditions were optimized and included both cell pellet washing and the presence of agents capable of slowing down the interaction with serum, allowing for the selection of the NPLyo protocol as the best candidate for WBC-labeling.
3.5. Assessment of NP Uptake from WBC
In order to confirm that the acquisition of radioactivity from WBC is due to NP uptake from cells and not to non-specific labeling, FITC labeled NPLyo (FITC-NPLyo) were prepared. The nanoparticles had similar physical-chemical features to the pristine NPLyo formulation, with an average diameter of 238.7 ± 5.0 nm, PI of 0.352 and ζ potential value of −21.1 ± 0.3 mV. The WBC-labeling (II approach procedure) was applied to FITC-NPLyo, and freshly isolated WBC were incubated with the particles. Confocal microscopy images were acquired on isolated and purified cells after incubation with FITC-NPLyo, indicating that the detected fluorescence is ascribable to uptaken particles (
Figure 5). Additionally, the intensity and number of fluorescence spots increased with longer incubation times, in agreement with what already observed in terms of radioactivity for gallium labeled WBC by means of NPLyo and confirming the specific radiolabeling due to NP uptake.
Since it is known that the size of NP greatly influences the speed of internalization and, although this speed also depends on the type of cell considered, nanoparticles with an average diameter of less than 100 nm generally have higher uptake rates than nanoparticles with diameters between 100 and 500 nm [
28]. For NP with 200–300 nm average diameter, similarly to the investigated CMC-based NP, the uptake studies are generally performed with longer incubation times, at least 1 h, and kinetics evaluations up to 8 or 24 h [
29]. Additionally, the uptake of NPLyo was time-dependent even in short time (up to 45 min), and such timing restrictions were necessary because of the fast
68Ga
3+ radioactivity decay.
FT-IR imaging can probe the cellular macromolecular components through their fingerprint characteristic bands [
30]. The main characteristic bands attributed to proteins, lipids, DNA, and polysaccharides [
31] of WCB are summarized in
Table 5, whereas NPLyo labeled WBC morphology and related vibrational mapping (set on ~1620 cm
−1 band, Amide I) are displayed in
Figure 6.
As shown in
Figure 7, the NPLyo labeled WBC preserve the main macromolecular composition during incubation. The reduction of the 45 min ratio may be attributed to the increased concentration of saccharides, related to the uptaking of the CMC-based NP. Similar analyses have been recently proposed to collect the spectral mapping of disease states in tissues or assess the in vitro drug-related cellular toxicity [
31]. Thus, the obtained results suggest a non-cytotoxicity effect of NPLyo, as already seen by trypan blue exclusion test, and for a preserved functionality of WBC.
4. Conclusions
This is the first report on the possible application of pharmaceutical grade CMC-based NP as a kit-type component for WBC-radiolabeling. Both investigated approaches were practically applicable and adaptable to the current guidelines followed for 99mTc-HMPAO WBC-labeling. The lyophilized powder approach is thought to be preferred, due to both handling aspects of labeling protocol and longer storage stability. In vitro investigations have confirmed the labeling of WBC due to NP uptake, without altering cell vitality.
NP marking was managed under mild conditions, using the acidic
68Ga
3+ 0.05 N HCl solution, freshly eluted on-site by the
68Ge/
68Ga generator. The direct use of the elution solution opens to the future possibility of transferring the investigated CMC-based NP to PET/CT operating also with longer-living radiopharmaceuticals, such as
44Sc, more suitable for the whole assessment of the WBC kinetic over time. Actually,
44Sc has a longer half-time (3.92 h) and can be produced either by
44Ti/
44Sc generator or by small cyclotrons, in both cases by acidic HCl solutions [
32]. Additionally, in order to accelerate WBC-labeling rate, the reduction of NP average diameter to less than 100 nm by upgrading to microfluidics apparatus [
33] or electrodynamic atomization [
34], and the covalent binding of specific chelators [
13] might be proposed.