Celia
Camacho-Toledano
ab,
Isabel
Machín-Díaz
ab,
Rafael
Lebrón-Galán‡
a,
Ankor
González-Mayorga§
c,
Francisco J.
Palomares
d,
María C.
Serrano
*d and
Diego
Clemente
*abe
aNeuroimmune-Repair Group, Hospital Nacional de Parapléjicos (HNP), SESCAM, Finca La Peraleda s/n, 45071-Toledo, Spain. E-mail: dclemente@sescam.jccm.es
bCentro de Investigación Biomédica en Red de Enfermedades Neurodegenerativas (CIBERNED), Carlos III Health Institute, Av. Monforte de Lemos, 3-5, 28029-Madrid, Spain
cLaboratory of Interfaces for Neural Repair, Hospital Nacional de Parapléjicos, SESCAM, Finca La Peraleda s/n, 45071- Toledo, Spain
dInstituto de Ciencia de Materiales de Madrid (ICMM), Consejo Superior de Investigaciones Científicas (CSIC), Calle Sor Juana Inés de la Cruz 3, 28049-Madrid, Spain. E-mail: mc.terradas@csic.es
eDesign and development of biomaterials for neural regeneration, HNP, Associated Unit to CSIC through ICMM, Finca La Peraleda s/n, 45071-Toledo, Spain
First published on 18th March 2024
Despite the pharmacological arsenal approved for Multiple Sclerosis (MS), there are treatment-reluctant patients for whom cell therapy appears as the only therapeutic alternative. Myeloid-derived suppressor cells (MDSCs) are immature cells of the innate immunity able to control the immune response and to promote oligodendroglial differentiation in the MS animal model experimental autoimmune encephalomyelitis (EAE). However, when isolated and cultured for cell therapy purposes, MDSCs lose their beneficial immunomodulatory properties. To prevent this important drawback, culture devices need to be designed so that MDSCs maintain a state of immaturity and immunosuppressive function similar to that exerted in the donor organism. With this aim, we select graphene oxide (GO) as a promising candidate as it has been described as a biocompatible nanomaterial with the capacity to biologically modulate different cell types, yet its immunoactive potential has been poorly explored to date. In this work, we have fabricated GO films with two distintive redox and roughness properties and explore their impact in MDSC culture right after isolation. Our results show that MDSCs isolated from immune organs of EAE mice maintain an immature phenotype and highly immunosuppressive activity on T lymphocytes after being cultured on highly-reduced GO films (rGO200) compared to those grown on conventional glass coverslips. This immunomodulation effect is depleted when MDSCs are exposed to slightly rougher and more oxidized GO substrates (rGO90), in which cells experience a significant reduction in cell size associated with the activation of apoptosis. Taken together, the exposure of MDSCs to GO substrates with different redox state and roughness is presented as a good strategy to control MDSC activity in vitro. The versatility of GO nanomaterials in regards to the impact of their physico-chemical properties in immunomodulation opens the door to their selective therapeutic potential for pathologies where MDSCs need to be enhanced (MS) or inhibited (cancer).
In the last years, the study of myeloid-derived suppressor cells (MDSCs), a highly immature regulatory cell type of the innate immune response, is gaining importance in the context of MS.7,8 MDSCs have been shown as key elements in the control of symptoms recovery in the murine MS model, experimental autoimmune encephalomyelitis (EAE).9 In fact, MDSCs are being identified as a target for new MS interventions due to their multifaceted roles as orchestrators of the control of the adaptive immune response (i.e., T lymphocytes),10–13 together with their involvement in favouring myelin restoration14 and preventing the harmful microbiota disbiosis induced by EAE.7In vivo studies in EAE have shown that the maintenance of the number and immature state of MDSCs is crucial for them to exert their potent control of T cell viability (immunosuppression) and moderate disease severity.9,13 However, once isolated from the donor organism, MDSCs tend to differentiate spontaneously, which greatly impairs their potential usefulness for future cell therapies requiring their isolation and prior expansion on standard cell culture conditions. For this reason, research is needed on new materials for cell growth able to preserve MDSCs as immature and highly immunocompetent cells.
Nanomedicine has emerged as a revolutionary field to provide novel customizable therapies in the nanoscale for a wide plethora of medical applications. Specifically, it makes use of nanosized tools for the diagnosis, prevention and treatment of diseases.15 The interest is such that an increasing number of applications and products containing nanomaterials, or at least with nano-based claims, have been made available to date. Within those, nanosized particle-based platforms for immune-related biomedical applications are a relatively novel field in which both immunosuppression and immune activation are being pursued, including immunotherapy tools, vaccine carriers, adjuvants, and drug delivery systems to target inflammatory cells. Among the most relevant nanomaterials under investigation for biomedical applications, graphene-derived materials (GDMs) are becoming promising candidates for both diagnostic and therapeutic uses. Regarding immune-related applications, the exploration of GDMs is limited and mainly focused on three major topics: (i) their interaction with immune cells for systemic biocompatibility assessment and the induction of specific immune responses, (ii) the development of immune-biosensors and (iii) their use in combination with antibodies for tumor targeting.16 In the first case, although largely unexplored until recently, GDMs are displaying a surprisingly attractive ability to interact with immune system elements, either by stimulating or suppressing specific responses. To this regard, their different physico-chemical properties including purity, shape dimension, redox state, and functionalization are essential drivers of this immune interaction. However, and despite discrete progress in the field, the role that these properties play on the specific interaction of GDMs with immune cells is still poorly understood.
Recent advances on the exploration of GDMs, including graphene oxide (GO), on immune cell interactions have shown important immunosuppressive actions over different myeloid cells,17 being more extensively explored in macrophages.18 Particularly, GO seems to have an outstanding ability to induce macrophage cytotoxicity, but also to alter their phagocytic capacity and, more importantly, polarize them towards either destructive or regenerative phenotypes by manipulating different physico-chemical properties of this nanomaterial such as its oxidation degree.19,20 Alternatively, oxidised multiwall carbon nanotubes, another class of GDMs, can be engulfed by dendritic cells (DCs) without altering the expression of specific markers such as those important for antigen presentation, i.e. major histocompatibility complex class II (MHCII), or the co-stimulatory molecule CD86, and cytokines such as TNF-α, but promote Th1 pro-inflammatory polarization of T lymphocytes by inhibiting the expression of anti-inflammatory Th2 cytokines.21 When further functionalized with a recombinant form of the human surfactant protein D, oxidised and carboxymethyl cellulose-coated carbon nanotubes are phagocytosed by macrophages/monocytic cell lines and increase pro-inflammatory destructive responses.22 Additionally, GDMs are able to modulate mesenchymal and neural stem cell differentiation in vitro,23 thus bringing an alternative therapeutic potential to these nanosized carbon materials for their use in cell-based treatments as culture devices to either favour or hamper cell differentiation.
Herein, we describe the pioneer use of GO-based culture substrates for the modulation of MDSC phenotype, specifically preventing the loss of their highly anti-proliferative activity (immunosuppression) over activated T lymphocytes in vitro in the context of the murine model of MS. Moreover, we identify that the oxidative degree and roughness of GO exerts a pivotal role on cell shape and differentiation, immunosuppressive activity, and viability.
Next, a careful chemical characterization was performed by X-ray photoelectron spectroscopy (XPS). Fig. 1D displays the comparison between X-ray XPS spectra of the C 1s core level of the GO slurry and rGO200. Spectra were normalized to the maximum intensity to highlight line shape differences, which provides direct valuable insights regarding the chemical environment of C atoms. The elemental content of C and O for the different samples is provided in Table 1, as well as the relative percentage of the different C–O functional groups. From the quantitative analysis of XPS peaks upon sample annealing, it is observed that the O versus C (O/C) atomic content ratio was dramatically reduced from 0.86 (GO slurry) to 0.10 (rGO200). This fact is also in line with the signal in the energy region of C 1s where the photoelectrons from C–OH and O–C–O groups are emitted, whose intensity also decreased with the O 1s signal (see below). When analyzed in detail, XPS spectra of GO exhibit its characteristic photoelectron emission comprised of two wide and intense peaks in the binding energy range from 282 to 290 eV. Detailed peak shape analysis was performed by the deconvolution of the C 1s spectrum with several Gaussian/Lorentzian symmetric components (ratio of 85/15) using a least-squares fitting routine. The energy position of the peaks and their relative heights were determined to account for the emission ascribed to the different chemical environment of carbon atoms according to the values reported in previous work.24,25 The result of this fit provides a single symmetric peak from C–C emission representative of the oxide nature of GO and several components attributed to the presence of various oxygen functional groups denoted as ∑(C–O) in Table 1. On the contrary, C 1s emission from rGO200 displays a very different lineshape dominated by an intense and asymmetric peak centered close to 284.5 eV. In addition, the long tail on the high energy side of the main peak usually reveals the metallic character of the sample, which certainly supports its graphitic-like nature and ascribes to sp2 bonding. Then, core level fitting of the spectrum is done by the deconvolution of an asymmetric component characteristic of C sp2 bonding (whose parameter values – peak position and full width half maximum – were obtained by fitting the C 1s signal from a HOPG reference sample) besides the symmetric components from the functional groups ∑(C–O). In this case, curve fitting of the whole spectrum makes it hard to clearly identify the sole emission from C–O and CO chemical states, so their full contribution is added in Table 1. Moreover, a weak component appears, broader that any of the previous ones, shifted at a higher binding energy value of ca. 6.5 eV from the main C 1s peak, which is clearly linked to the characteristic π–π* shake-up transition consistent with the majority existence of sp2 bonding in the sample. Note that a small amount of disordered regions where C atoms with defect-like sp2/sp3 bonding nature might also exist and those are included in the C–C component, but with no significant weight due to the low chemical shift and their minor contribution.
Sample | O/C (at) | C–C | Σ(C–O) | C sp2 | C–C | C–O | C![]() |
O–C![]() |
π–π* |
---|---|---|---|---|---|---|---|---|---|
GO | 0.86 | 42.4 | 57.6 | — | 42.4 | 47.9 | 3.8 | 5.9 | — |
rGO90 | 0.48 | 49.6 | 50.4 | — | 49.6 | 43.0 | 2.0 | 5.4 | — |
rGO200 | 0.10 | 76.5 | 23.5 | 75.3 | — | 18.0 | 5.5 | 1.2 |
In order to confirm the reproducibility of substrate preparation, we repeated AFM and XPS measurements in a representative sample of each batch prepared. Surface roughness was comparable in all batches tested for each particular substrate (Fig. S1†). Specifically, rGO200 samples turned out to be very similar intra- and inter-batches (e.g. Rq values differed less than 2.5% within each particular batch and less than 17% between batches). Statistical comparisons corroborated the absence of significant differences for samples from different batches (p = 0.458 for Rq, p = 0.450 for Ra and p = 0.384 for Rmax) and samples from the same batch (p = 0.876 for Rq, p = 0.791 for Ra and p = 0.694 for Rmax). XPS spectra analyses also confirmed reproducibility among batches with respect to chemical composition (Fig. S2†).
Encouraged by the reported implications of GDMs on cell differentiation,23 rGO200 films spin-coated on top of conventional glass coverslips were used to modulate the phenotype of splenic MDSCs in vitro (Fig. 2A). First, we confirmed that glass did not exert a noticeable alteration of MDSC differentiation after 24 h in culture compared to those grown on tissue culture plastic (treated polystyrene) except for an increase in the macrophage marker F4/80 (epidermal growth factor-like module-containing mucin-like hormone receptor-like 1) in the absence of the differentiation markers of DCs (the integrin alpha X, known as CD11c, their most widely used defining marker) and antigen presenting cells (APCs; marked with MHC-II, only found on cells with the ability of presenting antigens to T lymphocytes) (Fig. S3†). Mean fluorescence intensity (MFI) is a valuable technical approximation used in flow cytometry to define the extent of immunostaining and, in our particular case, cell differentiation, i.e. the higher the MFI of phenotype-defining markers, the more differentiated state of the immune cells tested. The absence of an increase of the MFI values for Ly-6G (typical marker of granulocytes) and CD11c (DCs) after culture on glass ruled out the possibility of a spontaneous differentiation of our MDSCs towards those two immune cell types in vitro.9 However, these same MDSCs on glass showed a significant increase of the macrophage marker (F4/80), which is also a specific marker of monocytic MDSCs.27 One of the main characteristics of MDSCs is their absence of antigen-presenting capacity measured by MHC-II expression (which differentiates them from pro-regenerative macrophages that present a mild expression of MHC-II28). Therefore, the negligible MHC-II immunolabelling found in these cells indicates that the increase in F4/80 found is not indicative of their maturation towards either pro-inflammatory (high expression of MHC-II) or pro-regenerative (low/intermediate expression of MHC-II) macrophages. Finally, CD11b (a panmarker of myeloid cells) and Ly-6C (marker for monocytic MDSCs), which have been both endorsed as markers for elevated immunosuppressive activity in MDSCs,13,29 did not show any significant alteration. Based on this, these data indicate that culture on glass does not induce a remarkable modification of the cell phenotype of these MDSCs compared to tissue culture plastic in terms of maturation and activation.
On the contrary, rGO200 substrates were able to maintain splenic MDSCs in a less matured state than those cultured in conventional glass coverslips at the same time point (Fig. 2B and C). Specifically, rGO200 substrates significantly prevented splenic MDSC differentiation towards mature macrophages (F4/80 in MFI), DCs (CD11c+; both in percentage of cells and MFI) and APCs (MHC-II+; both in percentage of cells and MFI).
In order to determine whether maintaining the undifferentiated state of splenic MDSCs after culture on rGO200 could have any impact on their immunosuppressive capacity, these cells were co-cultured (1:
4) with splenocytes obtained from EAE mice at the peak of their clinical course and previously stimulated with the myelin oligodendrocyte glycoprotein peptide (MOG35–55), the same molecule used for the immunization (induction) of the EAE model, (Fig. 2D). As previously showed for tissue culture plastic,13 splenic MDSCs cultured for 24 h on glass coverslips did not exert any immunosuppression on MOG-stimulated T lymphocytes, i.e. T cells that respond to the immunogenic stimulus used for active immunization (proliferation index (PI) vs. non-stimulated splenocytes, PI: MOG = 5.7 ± 0.8 and glass coverslip = 6.4 ± 0.8%; p = 0.516). Remarkably, splenic MDSCs exposed to rGO200 for 24 h were able to significantly suppress T cell proliferation (PI: rGO200 = 4.1 ± 0.5%, p < 0.05 vs. MOG; Fig. 2E). These data point to rGO200 films as substrates to potently preserve the immunosuppressive capacity of splenic MDSCs by keeping them in an undifferentiated state in culture.
In an attempt to discard that rGO200 nanosheets were being released from the culture substrates and exert a direct effect on T cells in suspension, culture media from the different conditions under investigation (but in absence of cells) were analysed by dynamic light scattering (DLS). We identified a limited population of particles smaller than 100 nm in Z average (Polydispersity Index, PDI = 0.2 and mean size = 10 nm). Comparatively, the hydrodynamic size of the original GO slurry in suspension was much higher (Zaverage = 5400 nm, PDI = 0.2 and mean size = 1350 nm) (Fig. S4†). Therefore, the observed decrease in T-cell proliferation after exposure of MDSCs to rGO200 should be attributed to a modification of cell activity by direct contact with the film rather than to the presence in suspension of rGO nanosheets.
After substrates characterization, the immunophenotype and immunosuppressive activity of MDSCs were addressed after cultured on rGO90versus rGO200. The ambitioned use of MDSCs as a cell-based therapy in autogenic transplants forced the use of the bone marrow instead of the spleen as the main tissue source for this regulatory cell type. For this reason, we next used bone marrow MDSCs instead of splenic MDSCs to facilitate the future translation of our findings. The exposition to either rGO90 or rGO200 substrates reduced both the cell percentage and MFI of CD11c+ (DCs) and MHC-II+ cells (APCs) in bone marrow MDSCs (Fig. 3A–D), similarly to results described for splenic MDSCs. Interestingly, rGO90 exposition dramatically diminished the percentage, not only the MFI value, of bone marrow MDSCs expressing the macrophage marker F4/80, not observed before for rGO200. Comparing both types of rGO substrates, rGO90 induced a more noticeable effect than rGO200 on the cell percentage and MFI value for both F4/80 and MHC-II markers (Fig. 3E and F).
To investigate how the culture on rGO200 and rGO90 films could impact the immunosuppressive activity of bone marrow MDSCs, cells were co-cultured with MOG-stimulated splenocytes isolated from EAE mice sacrificed at the peak of the clinical course. As for splenic MDSCs, these bone marrow MDSCs lost their immunosuppressive activity over antigen specific T cells when cultured on glass coverslips (PI vs. non-stimulated splenocytes: MOG-stimulated splenocytes = 8.6 ± 1.2; glass coverslip = 7.4 ± 1.2; p = 0.649). Contrarily, this immunosuppression was preserved when previously exposed to rGO200 films (PI: GO200 = 4.4 ± 1.0; p < 0.05 vs. MOG), but not when exposed to rGO90 (PI: rGO90 = 8.4 ± 2.0; p = 0.141 vs. MOG; Fig. 3G), despite the reduction in bone marrow MDSC differentiation markers also found for rGO90. These data indicate that bone marrow MDSCs cultured on rGO200 films, but not those on rGO90, retain the physiological ability to prevent the proliferation of T lymphocytes responsive to the immunogen causing the development of the EAE MS model.
Studies addressing the modulation of splenic MDSCs by the exposure to different biomaterials are very scarce in the literature, only circumscribed to cancer research and without analyzing each MDSC subset (polymorphonuclear or monocytic) separately. In this sense, immunosuppressive activity of MDSCs has been downregulated through their exposure to the DNA demethylating agent hydralazine (together with mitoxantrone as a chemotherapeutic agent) by means of a multifunctional nanoplatform.31 Moreover, repolarization of MDSCs toward a more pro-inflammatory phenotype has been also proved by using two cationic polymers: cationic dextran and polyethylenimine.32 In this context, our study pioneers to prove specific immunomodulation on the subset of monocytic MDSCs, when isolated from both the spleen and the bone marrow of mice with EAE.
Importantly, most of the studies carried out to date with GDMs and immune cells have been performed with nanomaterials in suspension, likely maximizing their impact on cell responses. Indeed, the exposure to nanomaterials in solution is expected to have a larger biological impact as the contact surface for interaction for both cells and nanomaterials is maximized, so internalization might be boosted. Our particular approach deals with rGO substrates in which rGO nanosheets are integrated in stable films attached to the underlying glass coverslips. In fact, our DLS data proved that drifting rGO nanosheets were almost absent in all culture conditions, thus ruling out any major biological responses mediated by rGO nanosheets in suspension. Interestingly, previous studies by others have demonstrated that GO films could induce cell proliferation and apoptosis by cell contact, independently of cell engulfment. For instance, Escudero et al. demonstrated that coating CoCr alloys with electrochemically reduced GO caused lower damage to cell plasma membranes in J774A.1 macrophages than pristine alloys.33 In a different study, the diminished GO internalization observed in bone marrow macrophages with a significant deficiency in phagocytosis did not affect cell necrosis induced by these GO nanosheets.34 In the same sense, previous work has shown that neither pristine nor functionalized graphene flakes induced any direct effect on human T-cell proliferation and suppression.35 Interestingly, a former study using an immune array of human peripheral blood mononuclear cells (PBMCs) provided convincing evidence that the adaptive immune response observed after GO exposure was not directly, but indirectly, affected by the primary effect on myeloid cells. Indeed, genes involved in direct T-cell activation, such as IL-2 and IFN-γ, were not affected, which correlated with the absence of CD69 and CD25 expression in these cells after GO exposure.36
In our study, GO films were early discarded as culture substrates because GO nanosheets without any thermal treatment got rapidly released from the films, thus causing a major biological impact of nanomaterials mediated by interactions in suspension rather than from the culture substrate. Thermal conditions for rGO90 were chosen as the ones sufficient both to significantly reduce GO nanosheets (O/C ratio from 0.86 to 0.48) and to preserve the integrity of the film without nanosheets release. Finally, those conditions to obtain rGO200 were designated again to produce a further significant reduction of rGO (O/C ratio of 0.10). Based on previous work by others, the oxygen-containing groups remaining in our rGO200 films would need much higher temperatures (>600 °C) to be effectively removed from the films.37 Indeed, our observations clearly indicate that annealing at 200 °C is an optimal option to maintain MDSCs in a highly active anti-inflammatory state. This observation is in agreement with the decreased oxidative stress and pro-inflammatory cytokine secretion observed in RAW-264.7 macrophages after exposure to suspensions of rGO nanostructures also annealed at 200 °C for 30 min.20 In both cases, the thermal reduction of GO at 200 °C improved the biocompatibility of this nanomaterial. Therefore, our data provide further insights into the possibility of triggering an anti-inflammatory response by GDMs, such as the site-specific Th2 response induction (i.e. a significant increase in IL-5, IL-3, IL-33, and their soluble receptor, sST2), as previously postulated by Wang et al. for T cells exposed to graphene nanosheets in suspension.38 Our data also indicate that, by modulating the physico-chemical characteristics of rGO culture substrates, we might exert either pro- or anti-inflammatory effects on bone marrow MDSC activity. These findings are in agreement with previous reports showing the ability of GDMs in suspension to induce changes on immune cell activity. For instance, small GO nanosheets were able to up-regulate critical genes involved in both Th1 and Th2 immune responses such as CSF2, TNF, IL6, IL10, CD80, IL1, IL1R1, TICAM1, IL8, IL23A, NFKB1, TBX21, CD40, CCR6, and IFNAR1 on human PBMCs. Also, this GO nanomaterial stimulated the release of some Th1 but also Th2 cytokines such as IL-1β, IL-1α, TNF-α, IL-10, IL-6, and IL-8 in the same cells.36
Our morphological studies by SEM showed that both rGO200 and rGO90 inhibit the spontaneous increase in cell size and protrusion elongation typical of myeloid cells in culture.30 In agreement with our data, Sasidharan et al. showed that pristine graphene flakes in suspension (fabricated by the arc discharge physical method) impeded myeloid cells such as RAW264.7 macrophages to attain their normal stretched morphology and the formation of filopodial extensions,35 this effect being significantly reverted by carboxyl functionalization of the flakes. In a different work, Hung et al. showed that GO significantly elicited dendritic and/or round-like morphologies in human monocytes as compared with the well-spread morphology of the control group (glass coverslips).39 Although our data point towards the direct physical interaction with rGO films in culture as responsible for this limitation in cell body spreading, we cannot exclude the possibility of an indirect effect of rGO. It has been previously shown that changes in myeloid cell morphology induced by GDMs are mediated in an autocrine manner.40 When cultured in control conditioned media collected from untreated RAW264.7 cells, naïve RAW264.7 cells frequently displayed a multipolar stellate morphology, typical for macrophage polarization. However, when exposed to conditioned media collected from graphene-exposed RAW264.7 cells, naïve RAW264.7 cells showed a polar spherical phenotype, likely mediated by soluble graphene-induced factors. Contrarily, macrophages grown on electrochemically reduced GO presented elongated cell bodies with characteristic protrusions.33 These controversial observations are probably related to the different physico-chemical properties of the GDMs under investigation, the cell source and phenotype, and the specific strategy followed for cell culture.
Our data also suggest that rGO90, but not rGO200, exerts cytotoxic activity on bone marrow MDSCs as its exposure induces both a reduction in the number of viable cells and an increase in those at a late stage of cell apoptosis. To this regard, the toxicity of GDMs on myeloid cells is still a matter of debate.41 Some authors have reported the so-called “mask effect” to explain cell toxicity induced by GDMs.42 Specifically, the close contact between GO nanosheets and cell membranes could drive to GO internalization. Once inside, GO could directly affect cell parameters such as viability, redox state and activation. The normal signalling and communication of cells with their environment could be also significantly altered by the presence of GO nanomaterials in their surroundings, thus influencing cell homeostasis and altering cell death, proliferation and activation cascades. In this sense, the smaller the size of GO sheets, the larger the mask effect expected. For instance, GO nanosheets (lateral dimension of 1–2 μm) in suspension were found to induce a toll-like receptor 4-mediated necrosis in different lines of macrophages, with a clear decrease on cell protrusions.34 When tested in vivo, subdermal inoculation of GO nanosheets in a rat model increased CD163+ macrophages (M2-like phenotype) compared to control animals, whereas no changes in CD86+ macrophages (M1-like) were detected.39 In sum, myeloid cells are able to engulf and process GDMs, markedly depending on physico-chemical properties such as shape, size, and chemical functionalization of the nanomaterials, which might either induce pro/anti-inflammatory responses or lead to the modulation and disruption of the immune activity.43
Other authors have also postulated nanomaterial concentration as an important factor for the control of the activity of MDSCs. In a particular study with MDSCs derived from PBMCs of healthy human donors, low concentrations of GO nanosheets in suspension (2.5 and 5 μg mL−1) induced MDSC proliferation, while a higher concentration (10 μg mL−1) reduced their viability.44 Other than these properties, chemical modifications (i.e. functionalization) of GDMs have also demonstrated to significantly impact their toxicity and ability to modulate myeloid cell interactions. For instance, polyethylene glycol (PEG) reduced the risk of immune responses of nanomaterials by increasing their stability in physiological conditions and minimizing their interaction with biomolecules. Particularly, Feito et al. described how PEGylated GO diminished the proliferation of macrophages in a concentration-dependent fashion.45 In combination with polyethylenimine to functionalize GO and serve as an adjuvant, PEG promoted DC maturation and enhanced cytokine secretion by acting over toll-like receptor-dependent pathways.46 However, no data are available to date about the effect of GO functionalization over MDSCs, a field that seems pivotal for its future therapeutic use in nanomedicine. It is important to note that other GDMs such as carbon nanotubes could also represent attractive immunomodulatory tools for MDSCs, based on their capacity to positively interact with immune cells such as DCs without altering the expression of some of their specific markers and to promote specific actions on the polarization of T lymphocytes.47,48 Moreover, oxidized multiwall carbon nanotubes initiate macrophage phagocytosis and upregulate CD14, CD11b, TLR-4/MD2, and CD206, without altering MHC-II expression. Macrophages activated with these nanomaterials produce angiogenesis-related cytokines such as MMP-9 and VEGF and reduce levels of pro-inflammatory cytokines.49 The understanding and precise control of the physico-chemical properties of these carbon-based nanomaterials is pivotal for exploiting their therapeutic potential.
In our studies, different thermal annealing conditions resulted in rGO films with slightly but significantly different roughness and redox states. These both physico-chemical parameters were closely dependent on each other based on the thermal treatment used, with the latter being expected to have a superior implication given its more substantial difference between rGO90 and rGO200 participating in the opposite modulation of immunological responses from endogenous monocytic MDSCs isolated from the bone marrow in an autoimmune disease. The specific molecular mechanisms behind the different modulation of MDSC behaviour driven by these two physico-chemically dissimilar rGO substrates will be the focus of future work in our laboratory. Nonetheless, we can hypothesize some pivotal actors contributing to these results based on previous findings in comparable systems. Extensive literature has already outlined the pivotal role that the biophysical cues of biomaterials including nanomaterials play in the manipulation of (stem) cell fates.50 Based on previous work by others, we theorize that the MDSC affectation mediated by rGO90 is likely related to its higher oxygen content in comparison to rGO200. Although still not investigated for MDSCs, published data on neural cells consistently indicate induction of reactive oxygen species (ROS) production, primarily through activation of mitochondria and NADPH oxidases located in the cytoplasm and plasma membrane. These events are followed by lipid peroxidation, cell membrane damage and ROS-disrupted mitochondrial homeostasis associated to the exposure to more oxidized forms of GO nanomaterials in suspension.51 In these studies, the autophagy-lysosomal network was proved to be initiated as a defensive reaction to eradicate oxidative damage to mitochondria, whose failure due to lysosomal dysfunction exacerbated mitochondrial stress and led to cell apoptosis/necrosis. In a different work with a human monocyte cell line (THP-1 cells), a reduction in GO was directly associated with a lower cytotoxicity; while more oxidized GO forms led to increased lipid peroxidation, membrane leakage and cell death.52 In agreement with this, our rGO90 films (with a higher content in oxygen-containing groups as proved by XPS studies) promoted lower viability and higher late apoptotic percentages in comparison to glass and rGO200 substrates.
In our studies, different thermal annealing conditions of GO coatings resulted in rGO films with slightly but significantly different roughness and redox states. These both physico-chemical parameters are therefore closely dependent on each other in this model, being difficult to separate the respective contribution of each one on MDSC responses. Nonetheless, given the more dramatic changes in O/C ratios in comparison to roughness values, we hypothesize the former as a more relevant parameter for the effects found on MDSCs growth, differentiation state and immune functionality.
Thermal treatment of rGO materials typically induces the removal of oxygen-containing groups and the recovery of sp3/sp2 configurations and π–π transitions coming from graphitic structures.37 As a consequence of these chemical changes, there is a certain recovery of the original stacking of the graphene sheets (no longer repulsed from each other by the oxygen-containing groups) that could drive to a more effective disposition of the sheets and, therefore, a lower roughness of the surface of the resulting films. Surface roughness has been also proved to decisively impact cell fate. For instance, scaffold roughness values around 70 nm promoted osteogenic gene expression in mesenchymal stem cells, while those around 15 nm enhanced chondrogenic gene expression of this same cell type.53 Additionally, smooth glass surfaces (Rq = 1 nm) supported rapid proliferation and long-term self-renewal of human embryonic stem cells, while rough glass surfaces at the nanoscale (Rq = 150 nm) induced their spontaneous differentiation.53 Further work has even identified a roughness threshold for controlling pluripotency of embryonic stem cells (Rq < 392 nm for long-term maintenance of pluripotency; Rq > 573 nm for faster unidirectional differentiation).54 Importantly, this type of behaviour has been demonstrated to be cell-type specific, so precise conditions must be defined case by case. Moreover, although extensively explored for a diversity of cells including stem cells, immune cells have been scarcely explored in these scenarios, so conclusions must be taken carefully. For instance, differences between GO and rGO nanosheets have been explored in terms of cell toxicity, morphology and immunomodulation of bone marrow-derived macrophages with rGO being more cytotoxic than GO, with a marked reduction on cell complexity and exhibiting a more pro-inflammatory activity.55 Although these cell-specific effects exerted by rGO were concomitant with a substantial modification of nanosheet morphology (i.e., from a sheet-like structure for GO to a curly polygonal shape for rGO), its significance on cellular activity and the mechanisms underlying this impact on immune cells were not further explored. In a different study, it has been also reported that thermally-reduced GO sheets promoted RAW-264.7 macrophages polarization towards a more anti-inflammatory activity state.38 In such work, although the thermal reduction of GO provoked a more crumpling surface morphology of the tested sheets compared to the flat-like GO, the direct effect of this morphological change on macrophage cell activity was not explored. rGO200 films are herein presented as attractive substrates for the preservation of the immature phenotype of MDSCs in culture, with the consequent potential for their use as a cell therapy for immune diseases. Alternatively, those based on rGO90 films could be of interest for the implementation of novel therapeutics in cancer, including the incorporation of stimuli-responsive nanomedicines with a mediated activation by external stimuli such as light and radiation.56
AFM studies were carried out by using a Nanoscope V forces microscope (Bruker) in tapping mode with a TESPSS tip (Bruker). Surface roughness parameters (i.e. Rq, Ra and Rmax) were obtained by using the Nanoscope Analysis software v2.0. Rq is the root-mean-square (rms) value of the profile heights over the evaluation length. Ra is defined as the arithmetic average of the profile heights over the evaluation length. Rmax is the largest of the successive values of the vertical distance between the highest and lowest points of the profile calculated over the evaluation length.
Culture media after incubation with both types of rGO culture substrates were analyzed by DLS for the detection of rGO nanosheets released from the substrates. Measurements were carried out in a ZetasizerNano ZS instrument (Malvern). Temperature was set to 25 °C and samples were diluted with MilliQ distilled water. Data analysis was performed considering the Gaussian distributions intensity-weighted and numbered weighted, obtaining the Z average and PDI from the first, and the mean particle size in number from the second one.
To obtain MDSCs from bone marrow, mice were euthanized at the peak of their clinical symptoms defined as for splenic MDSCs. Both femurs and tibiae were dissected out and the bone marrow was flushed with supplemented RPMI media. Bone marrow cells were centrifuged at 437g for 5 min and the obtained pellet was resuspended and red cell removed with ACK lysis buffer. After that, BM-derived cells were centrifuged at 437g for 5 min, resuspended in sPBS, and passed through a 100 μm filter (BD Biosciences).
Splenocytes or bone marrow cells were resuspended in sPBS and the Fc receptors were blocked for 10 min at 4 °C with 10 μg mL−1 anti-CD16/CD32 antibodies (BD Biosciences 553142). After blocking, FITC conjugated rat anti-mouse Ly-6C (AL-21 clone), PE conjugated rat anti-mouse Ly-6G (1A8 clone) and PerCP-Cy5.5 conjugated rat anti-mouse CD11b (M1/70 clone) antibodies were added to the cell suspension and incubated for 30 min at 4 °C in the dark (Table S1†). The samples were then rinsed with sPBS, centrifuged at 371g for 5 min, resuspended again in sPBS, and sorted in a Fluorescence Activated Cell Sorter (FACS) Aria (BD Biosciences) located at the Flow Cytometry Service of the Hospital Nacional de Parapléjicos. Monocytic splenic and bone marrow MDSCs were identified as Ly-6Chigh Ly-6G−/low gated on CD11b+ myeloid cells and recovered at a purity >95%.
For cell viability analysis, labelled bone marrow MDSCs were stained with the Annexin V-FITC Apoptosis Detection Kit (Beckman Coulter, IL, Italy) following manufacturer's protocol. Briefly, cells were centrifuged at 93g for 3 min, rinsed in PBS and centrifuged again to remove supernatants. Pellets were resuspended in 100 μl of binding buffer at a concentration of 106 cells per mL and stained with 5 μl of AV-FITC conjugate plus 5 μl of PrI for 15 min at room temperature. The different cell staining patterns were classified as follows: (1) viable cells: double negative for AV and PrI (AV−PrI−); (2) early apoptotic cells: positive for AV and negative for PrI (AV+PrI−), and (3) late apoptotic/dead cells: double positive for AV and PrI (AV+PrI+). In all cases, MDSCs were analyzed in a FACS Canto II cytometer (BD Biosciences). The data obtained were analyzed using the FlowJo 10.6.2 software (Tree Star Inc.).
Footnotes |
† Electronic supplementary information (ESI) available: Data about used antibodies for flow cytometry, XPS, AFM and DLS. See DOI: https://doi.org/10.1039/d3nr05351b |
‡ Present address: Hospital Universitario de Toledo, Avd. Río Guadiana, s/n 45004. Toledo, Spain. |
§ Present address: Hospital Universitario de Navarra, C/Irunlarrea 3, 31008 Pamplona, Navarra, Spain. |
This journal is © The Royal Society of Chemistry 2024 |