Impact of Locally Administered Carboxydextran-coated Superparamagnetic Iron Nanoparticles on Cellular Immune Function

Interstitially-administered iron oxide particles are currently used for inter-operative localization of sentinel lymph nodes (SLNs) in cancer staging. Several studies have described concerns regarding the cellular accumulation of iron oxide nanoparticles relating them to phenotype and function deregulation of macrophages, impairing their ability to mount an appropriate immune response once an insult is present. This study aimed to address what phenotypic and functional changes occur during lymphatic transit and accumulation of these particles. Data shows that 60 nm carboxydextran-coated iron nanoparticles use a non-cellular mechanism to reach the draining LN and that their accumulation in macrophages induces transient phenotypic and functional changes. Nevertheless, macrophages recover their baseline levels of response within 7 days, and are still able to mount an appropriate response to bacterially-induced inflammation.


Introduction
The use of superparamagnetic iron oxide (SPIO) nanoparticles in biomedical applications has been the subject of extensive and growing research in the past 10 years. SPIO js970@mrc-cu.cam.ac.uk.

Competing Interests
Quentin Harmer is the CTO at Endomagnetics Ltd. and Eric Mayes is the CEO at Endomagnetics Ltd.. The other authors declare that they do not have any conflict of interest. nanoparticles are attractive for clinical use because they can form stable biocompatible, colloidal suspensions when coated, they are only magnetic in the presence of an applied field, they can be conjugated with a range of ligands and molecules, and they can be broken down by the body. This combination of features has resulted in proposed applications including imaging, sensing, targeted drug delivery, therapeutic heating (hyperthermia) and mechanical stimulation. [1,2] Much of this research has yet to be translated out of the laboratory, and iron oxide nanoparticle products approved for clinical use are more limited. [3][4][5][6] SPIO nanoparticles were first approved for use as intravenous MRI contrast agents due to their strong T1 and T2 relaxivities [7], and later as an iron replacement therapy. More recently, iron oxide nanoparticle suspensions have been approved for LN mapping and detection as part of cancer staging. [8] In this application, a colloidal suspension of SPIO particles is injected locally into interstitial tissue near the tumor. The particles are sized such that they are taken up by the lymphatics [9] and migrate to the draining 'sentinel' LNs where they accumulate. The magnetized nodes can be detected and localized during surgery using a handheld magnetic probe. [10,11] This sentinel node biopsy procedure is an important part of staging a number of cancers, particularly breast cancer and malignant melanoma, and the magnetic technique has been proved to be a safe and effective alternative to the previous standard technique that uses a radio-labeled tracer and a handheld gamma probe to detect the sentinel nodes. [11][12][13][14] The fate of intravenously administered SPIO nanoparticles for MRI and iron replacement is well understood: they are rapidly taken up by Kupffer cells in the liver and other cells of the mononuclear phagocyte system, after which they are metabolized and regulated by normal physiological iron homeostatic mechanisms. [15] However, the mechanism of uptake and the fate of locally-delivered SPIO nanoparticles are less well described. In breast cancer patients, it has been shown that locally-delivered SPIO rapidly accumulates in sinuses and the subcapsular space of SLNs [13], but the mechanisms by which SPIO was transported to the draining LNs and longer-term effects have not been investigated.
The size and properties of particles can determine the rate at which they transit by mechanical means to LNs [16][17][18][19][20][21][22][23], ranging from seconds to a day [17,24,25], with optimal transport via lymphatics achieved in particles between 20 and 60nm in diameter. In contrast, cellular transport via loaded antigen-presenting cells takes from 2 to 24 hours to traffic from the peripheral tissue to the lymph node (LN) [25][26][27][28][29][30][31], and subsequent retention of particulates in the LNs depends on size [32][33][34] and retention by phagocytic cells. [35,36] Initial analysis of injected nanoparticles in in vivo models and cancer patients indicated that, longer-term, accumulation in tissue macrophages was possible following scavenging at the injection site or respective draining LN. This raises the potential concern that accumulation of iron oxide nanoparticles may induce phenotype and function deregulation of macrophages impairing their ability to mount an appropriate immune response once an insult is present.
A growing number of studies have investigated cellular changes brought about by uptake of engineered SPIO nanoparticles, however results varied dramatically between cell types. Also, particles examined had different sizes (which can differ between 20 and 150nm) and different surface coatings [37] (which influence particle biodistribution and cellular uptake), therefore inducing different cellular responses. In non-immune neuronal cells, uptake of dextran-coated SPIO nanoparticles induced changes in the expression of a large number of genes, including genes related to iron homeostasis [38], albeit transiently. In contrast, these SPIO nanoparticles had no impact on leukocyte proliferation but induced slight immunesuppressor effects (such as inducing expression of IL-10 and members of the TNF family of proteins), as well as reducing monocyte cell migration. [37,39] In macrophages, carbohydrate-coated SPIO nanoparticles induced pro-inflammatory (M1) responses, leading to inhibition of tumor growth in mammary and lung cancer metastasis in the liver and lungs. [40] Whereas SPIO nanoparticles coated with dimercapto-succinic acid (DMSA), 3aminopropyl-triethoxysilane (APS) or aminodextran (AD) exhibited no effect in proliferation but induce changes in iron metabolism, increases in ROS, and induction of antiinflammatory (M2) phenotypes with reduced macrophage random migration and concurrent increased chemotaxis. [41] The latter was also a feature for citrate and malate-coated SPIO nanoparticles. [42] Carboxydextran-coated SPIO nanoparticles induced increases in endosomal recycling in human mesenchymal stem cells [43] and pushed M2 macrophages towards a M1-like phenotype. [44] Deregulation of macrophage activation from an M1 to a M2-like activation state upon LPS (bacterial lipopolysaccharide) inflammatory stimulus was also described, characterized by suppression of IL-10 production, enhanced TNF-alpha production, and diminished phagocytic ability. [37] However, the majority of these studies focus on immediate, short-term effects after exposure to SPIO nanoparticles and do not address longer-term effects of SPIOs. Thus, we sought to address the kinetics and means carboxydextran-coated SPIO nanoparticles use to reach the draining LN, and potential longterm implications of its accumulation, both in the injection site and the downstream LN. Here, we show that carboxydextran-coated SPIO nanoparticles rapidly transit to draining LNs via mechanical transport rather than via cell-dependent means. While particles persist at injection sites and LNs where they may be sampled by macrophages, no long-term impact on macrophage phenotype or functional capacity were observed in vitro or in vivo.

Mechanisms of SPIO nanoparticle transport to lymph nodes
To determine the means by which carboxydextran-coated SPIO nanoparticles reach the lymph node (LN) i.e. mechanical vs. cellular transport, a time course assay was performed in which a SPIO nanoparticle suspension was administered in conjunction with TRITC paint. TRITC painting gives a measure of cell migration from peripheral tissues to LNs, where peripheral antigen presenting cells are stimulated by the sensitizing agents or "danger signal" within the TRITC paint, which they engulf, and carry to the LNs. On a gross scale, it was immediately apparent that SPIO nanoparticles had reached LNs at the earliest time points examined (10 minutes post-administration). In contrast to non-draining nodes, which were pearlescent white, a brown signal was observed in the capsular regions of draining nodes (Figure 1a). Confocal analyses confirmed that SPIO rapidly reached draining LNs and could be detected by its carboxydextran coating within 10 minutes of inoculation ( Figure  1b), but no TRITC positive cells were present. At 10 minutes, SPIO were found predominantly in the subcapsular sinus, co-localized with lymphatic structures (marked by the lymphatic-specific glycoprotein LYVE-1) and to a lesser extent, deeper lymphatic structures of the medullary sinuses ( Figure 1b). While SPIO was predominantly localized to LYVE-1 + lymphatics, some association with CD11b/F480/sialoadhesin-expressing macrophages that line the sub-capsular sinus in close proximity to vessels was observed ( Figure S1a, left panel). By 24 hours, SPIO was still observed in the subcapsular sinus but was also detected associated with deeper medullary lymphatics ( Figure S1b). In contrast to early time points, by 24 hours, abundant TRITC positive cells were detected in the cortex of LNs (Figure 1b, right panel). The kinetics observed were consistent with SPIO possessing the optimal physical properties required for lymphatic-specific uptake and subsequent transport via lymph flow to LN. [16][17][18][19][20][21][22][23] Further characterization of LNs by flow cytometry demonstrated that levels of cellassociated SPIO remained at consistent low levels throughout all time points tested ( Figure  1c). Such low detection levels indicated that the majority of SPIO identified in LNs was indeed "free" within circulating lymph rather than cellular. In contrast, TRITC positive cells i.e. those that engulfed matter in the peripheral dermis and then migrated, were only detected at the 24-hour time point (Figure 1d, e). The discrepancy in LN homing kinetics between SPIO vs. cell transport (10 minutes vs. 24 hours respectively) together with the lack of colocalization of dextran and TRITC-labelled cells supported a mechanical mechanism governed by normal lymphatic drainage function over cell-mediated methods as the primary mode of particle transport for the SPIOs. From initial FACS analysis, it remained unclear as to whether the small amounts of cellular SPIO detected on CD45+ cells were "associated" with the cell surface or had in fact been internalized. To address this, staining cell surfaceassociated SPIO was followed by a permeabilization step before a second antibody incubation. This would allow detection of total cellular SPIO. Flow cytometric analysis confirmed no significant increases in SPIO immune cells upon permeabilization (Figure 1c), consistent with the notion that SPIO was not bound and internalized for degradation or antigenic sampling. While we did observe some localization of SPIO with macrophages, more so at later time points, we depleted macrophages using liposomal clodronate (Clodrosome) prior to SPIO injections to determine if macrophages contributed to SPIO transport and localization (Figure 1a), but no change in the drainage or accumulation kinetics was observed.
To address the possibility of delayed arrival of TRITC-labelled cell to LNs as a consequence of the requirement for TRITC and sensitizing agents to permeate the epidermis and be engulfed, dendritic cells (DCs) were isolated from age-matched C57BL/6 mice, live-labelled and co-injected sub-cutaneously with SPIO to directly compare the kinetics of dendritic cell and SPIO homing. As described earlier, SPIO were observed within the LNs at the earliest time points measured and were still apparent 24 hours later. In contrast, and consistent with the earlier TRITC painting data, labelled DCs could only be detected within LN 24 hours after inoculation ( Figure S1c). This was consistent with earlier TRITC paint data and indicated that rapid SPIO arrival at the SLN was independent of cellular processes, and was not a result of DC trafficking missed or underestimated due to limitations of the technique.

Long-term SPIO dynamics
Having measured the acute kinetics of SPIO transport and localization in draining LNs we next sought to characterize longer-term localization and accumulation both at the injection site (ears, Figure 2a) and LNs ( Figure 2b). Immediately post-injection, the carboxydextran coating of SPIO was observed diffusely within the tissue consistent with free solution from the injection bolus. Similarly, iron of the particle core was also dispersed throughout the injection site (Figure 2a, blue, right panel). Seven days post injection, the carboxydextran coating was still detectable within tissue, but to a lesser extent than T0 (Figure 2a, left panel), however, iron remained diffuse within the dermis. By 1 month, carboxydextran was no longer detectable at the injection site, although iron was still clearly identifiable. This remained the case even one year after injection ( Figure 2, right panels) indicating that longterm, iron was retained within resident phagocytic cells once the coating had been digested. This was also true in LNs draining the injection site, with more punctate, cellular patterns of localization still observed at 1 year ( Figure 2b, Figure S1a). However, whilst iron remained visually abundant within tissues, quantification of levels using the Sentimag probe showed that relative iron content was indeed decreasing with time ( Figure 2c). During this long-term analysis we assumed that particles likely accumulated in macrophage rich areas (Figure 2a, b).

Impact of SPIO on macrophage cell line phenotypes
Given that macrophages are important components of the innate immune system, we proceeded to determine the potential impact of SPIO accumulation on macrophage phenotype and function in vitro, using two widely used macrophage cell lines: J774.2 and RAW264.7. First, we examined the uptake of nanoparticles ( Figure 3a and Figure S2a, b and S3a) using two different iron concentrations for the assays: 0.1 mg/mL and 1 mg/mL, where the macrophages were saturated ( Figure S2a, b). As a comparison, 60 nm gold nanoparticles were used. As expected, both cell lines readily engulfed SPIO particles, and an increasing cargo was accompanied by an increase in the macrophage SSC-A as measured by FACS ( Figure 3a and Figure S2a, b and S3a) and imaging ( Figure 3b and Figure S3b). To recapitulate the acute vs. persistent effects that these particles may have in the macrophages, cells were analyzed 1 and 7 days after incubation with SPIO. By day 7, the SSC-A of cells returned to baseline, which implied that macrophages either secrete digested SPIO (iron aggregates, which we did not observe) or they simply divide their iron content between daughter cells.
We next sought to assess the influence of SPIO on proliferation, apoptosis, migration, and polarization in vitro ( Figure 3), since functional defects in homing and immune responsiveness resulting from long-term accumulation may be clinically detrimental. When saturated with iron particles, a decrease in proliferation over 96 hours was observed, ( Figure  3c and Figure S3c) and this was accompanied by a transient increase in apoptosis, which returned to baseline levels by day 7 (Figure 3d and Figure S2d). Similar behavior has been described for Kupffer cells upon uptake of Ferucarbotran SPIO particles. [45] In contrast, 0.1mg/mL had no significant impact on proliferation or cell death. An important note here was that the lower iron concentration used in vitro was still 56 times greater than the dosage used in the clinical setting. Twenty-four hours after uptake of SPIO particles, a decrease in the random migration parameters was observed with increasing concentration (Figure 3e and Figure S3e), but this was not unexpected since loaded cells were larger, they would require time to process their cargo and also require more energy to drag payloads. In light of recent studies that reported a shift in macrophage polarization upon iron nanoparticle uptake [40,41,46], we sought to determine whether SPIO impacted polarization, and potentially determine changes in downstream immune function, examining typical surface markers characteristic of pro-inflammatory M1 (CD86, PD-L1, MHCII) and immunosuppressive M2 (Arginase 1, CD163, CD206/Mannose receptor) phenotypes. Earlier reports indicated that nanoparticles transiently shifted M2 macrophages towards an M1-like phenotype in both disease free [41] and tumor settings. [40,46] Cells saturated with SPIO particles displayed a transient decrease in MHCII and CD163 returning to control levels by day 7, while both concentrations of SPIO reduced arginase levels ( Figure 3f and Figure S3f). Interestingly, gold nanoparticles induced CD86 and MHC levels indicative of maturation and activation. Although we observed a transient reduction in M2-associated markers, no clear shift in phenotype towards M1 was detected ( Figure S2c). We also analyzed the effect of the low molecular weight carboxydextran-coating of SPIO nanoparticles and saw that it alone was sufficient promote similar transient shifts in surface markers on both cell lines tested ( Figure S2d, e). In contrast to previous reports that focused on acute effects of particles on macrophage behavior, we also evaluated the longer-term effects. And while these indicated some immediate effects, changes were transient with macrophages recovering normal behavior by day 7. This was consistent with data showing skin macrophage populations renew every 4 to 9 days [47,48], and could account for the "dilution" of any initial effects.

Impact of SPIO on macrophage cell line function in vitro
While surface marker expression provides some indication of responses to particle uptake, functional outputs are more informative when considering potential long-term clinical impact. Thus, we examined characteristics including phagocytosis, antigen-presentation and cytokine production. Phagocytosis capacity was quantified using 1-micron Fluoresbrite ® fluorescent microspheres (to mimic bacteria or debris, Figure 4a) before and after incubation with SPIO. Again, we observed a transient decrease in phagocytic ability when macrophage cell lines received high concentrations of SPIO nanoparticles (Figure 4b and Figure S4a) which was not surprising since cells were saturated and at full capacity. Having tested different concentrations of iron, cells were unable to uptake more iron in the presence of concentrations above 1 mg/mL ( Figure S2a). Phagocytosis returned to control levels at later time points. Interestingly, the gold nanoparticles induced no change in bead uptake, which could reflect that macrophages were not saturated with the gold particles. We then sought to determine capacity to uptake and process antigen in the forms of TRITC-labelled ovalbumin (OVA) and DQ-OVA. An initial decrease in the uptake of TRIC-OVA was recorded compared with control, but this recovered by day 7 (Figure 4c and Figure S4b). Despite a significant reduction in antigen uptake of approximately 80% shortly after incubation, proteolytic processing was only reduced by 30% with 0.1mg/mL, indicating that although sampling may have been transiently impaired, the downstream processing pathways to antigen presentation remained intact Figure 4d and Figure S4c). Previous studies considered the effects SPIO in the absence of an inflammation insult. [38,41,[44][45][46] However, due to the importance of macrophage functionality during inflammation we also induced inflammation after incubation with SPIO to determine if the particles had any impact on responses. Because macrophages can modulate immune responses through their secretome we analyzed cytokine production of both cell lines using a LEGENDplex multi-analyte flow assay after SPIO uptake before and after an immune insult with LPS ( Figure 5 and Figure S5). The vast majority of the cytokines examined showed no significant differences under baseline conditions. The addition of LPS induced differential production of some cytokines (red boxes), although again, with the exception of IL-1 and IL-10 it was inflammation rather than the presence of particles that induced these changes. This would imply that particle-loaded cells were still capable of responding to inflammatory insults. Interestingly, the increase observed at late time points with IL-1 and concurrent decrease in IL-10 indicated a more functionally activated phenotype. At the same time, an increase in IL-27 with increasing payload may have been indicative of an attempt to boost IL-10 and regulate pro-inflammatory cytokine production. [49][50][51] We did not see an SPIO specific induction of M1 or M2 signatures following LPS induction of inflammation. Further analysis of polarization and phagocytosis capacity was also consistent with our previous analysis (Figure 3 and 4, and Figure S3 and S4) in the presence or absence of an inflammation insult ( Figure S6 and S7).

Impact of SPIO on macrophage function in vivo
As the doses in vitro far exceeded those experienced in vivo, even at the lowest concentration tested, we analyzed the impact of SPIO particles in macrophages in vivo at both the injection site (receiving a higher dose) or in the draining LN over time. There were no differences in immune cell populations at either site following inoculation (Figure 6a, b), but we detected dextran + macrophages at the injection site that had taken up nanoparticles immediately following injection (Figure 6b). This was not the case at the LN, or at later time-points, consistent with previous data examining transport kinetics and dextran vs. iron retention (Figure 1 and 2). At the draining LN, no significant alterations in macrophage phenotype were measured at any time point examined following SPIO injection (Figure 6c). At the injection site, decreased levels of PD-L1, Arg1, CD163 and CD208 were observed, as with in vitro assays, but these returned to levels comparable with controls by day 8 ( Figure   6d) indicating no long-term changes in macrophages at either site following SPIO administration.

Conclusion
Carboxydextran-coated superparamagnetic iron oxide particles (SPIO) have recently been implemented as reliable alternatives to radioisotopes for intraoperative mapping and identification of sentinel LNs in cancer staging. [8,[10][11][12][13] However, the mechanisms of transport to LNs and longer-term impact of tissue accumulation remained unclear. Here we identified that rapid transport to LNs was governed by mechanical rather than cell-mediated means [25][26][27][28][29][30][31] with SPIO detected in draining LN 10 minutes post-administration, and localizing predominantly with lymphatic structures. In contrast, cell-driven trafficking to lymph nodes required 24 hours. We also demonstrated that, consistent with previous reports,

Europe PMC Funders Author Manuscripts
Europe PMC Funders Author Manuscripts [37,[39][40][41][42][43][44] alterations in macrophage behavior could be observed immediately after exposure. However, these changes were transient, resolving to baseline levels by day 7. Moreover, macrophages retained the capacity to proteolytically process antigen and respond to an inflammatory stimulus following exposure to SPIO. [52] Thus, these data provide evidence to suggest that after an initial perturbation, carboxydextran-coated SPIO nanoparticles persisting within macrophages conferred no long-term disruption to macrophage phenotype or function.

Experimental Section
In vivo time-course for lymph node drainage analysis: TRITC Paint vs SPIO, and in vivo iron measurements

All in vivo experiments were performed with UK Home Office authority under Project
Licenses 80/2574 and P88378375. Tetramethylrhodamine-5-(and 6)-isothiocyanate (TRITC) and SPIO (Magtrace ® , Endomagnetics, Ltd, previously branded as Sienna +® ) solutions were prepared under sterile conditions. 10mg TRITC was diluted in 1:1 mixture of acetone and dibutylphthalate (500 μL of each) to give final concentration of 10 mg/mL. Immune competent C57BL/6 mice were anaesthetized using 150 μL anesthetic stock solution (stock: 0.5 mL (50mg) ketamine + 0.25 mL (5mg) xylazine + 4.25 mL saline. Once asleep, any hair was removed using clippers. Then 25 μL of TRITC solution was slowly applied to the skin surface and gently spread using the pipette tip. Immediately following TRITC application, 10 μL of SPIO was injected sub-cutaneously to each site. Cages were place in a temperaturecontrolled Vetbed during the recovery period. Applications of TRITC and SPIO were performed for time points of 10 minutes, 30 minutes, 1 hour, 2 hours and 24 hours. After 10 minutes, animals were euthanized using CO2 followed by cervical dislocation. Draining LNs (brachial, for shoulder injections, and superficial cervical, for ear injections) were transferred to tubes containing sterile PBS for cell isolation and FACS analysis or snap frozen in OCT cryo preservative using dry ice for tissue sections. Non-draining inguinal nodes were also collected. These steps were repeated for all other time points. For long-term tracing, SPIO was injected on mice ears and iron content monitored for a year using a handheld probe (Sentimag, Endomagnetics Ltd., Cambridge, UK). To, note that a human dose of SPIO is on average 32 μmol iron/Kg (this was calculated considering that humans get 112 mg of iron per injection and considering an average body weight of 62 Kg). For these animal experiments, a 250 μmol iron/Kg dose was applied, over 7.8 times higher than in humans (considering an average mouse weight of 20 mg).

In vivo macrophage depletion
5 days prior to the time course assay 200 μL Clodrosome liposome suspension (liposomal formulation of clodronate, Encapsula Nanosciences) or control liposomes were injected via the intra-peritoneal route into immune competent C57BL/6 mice. This was repeated 3 days later. Macrophage depletion was confirmed by flow cytometry. On day 0, SPIO was administered as described above, and LNs were collected at time points of 10 minutes, 1 hour and 24 hours.

Isolation and injection of pre-labelled dendritic cells
CD11c+ dendritic cells were harvested from spleens. Briefly, under sterile conditions, spleens were gently teased apart with 25G needles and enzymatically dissociated with 1 mg/mL collagenase D (Roche). Following incubation at 37°C for 30 minutes the enzymatic reaction was stopped with EDTA. The resulting cell suspension was passed through 70 μm nylon mesh to remove large debris. Erythrocytes were then removed by treatment with ammonium chloride solution (0.15 M NH 4 Cl, 1 mM KHCO 3 , 0.1 mM EDTA) for 5 minutes at room temperature. After centrifugation at 1200 rpm for 6 minutes, the splenocyte cell pellet was re-suspended in MACS buffer (PBS containing 2mM EDTA and 0.5% BSA). CD11c+ dendritic cells were then positively selected using CD11c+ magnetic microbeads (Mitenyi Biotec) according to manufacturer's guidelines. Once collected, dendritic cells were livelabelled. Cells were incubated in 1 μM CellTrackerTM Red CMTPX (Molecular Probes) for 10 minutes at 37°C and were then allowed to recover for 30 minutes in full RPMI media supplemented with 10% FBS and 1% penicillin/streptomycin. Cells were counted and pelleted. Purity and viability of cells was confirmed by flow cytometry. Upon successful anaesthetization of mice, the cells were mixed with SPIO working solution and subcutaneous co-injection was performed as described above.

Immunofluorescence
Harvested LNs were embedded in OCT cryo-preservative and sliced into 10 μm sections using a cryostat. Sections were air dried and stored at -80°C and were then subjected to standard immunofluorescence protocols. Briefly, slides were fixed in ice-cold acetone for 2 minutes at -20°C, then washed in PBS. This was followed by a blocking step, incubating samples in 10% chicken serum in PBS/0.5% BSA for 30 minutes, room temperature. Sections were incubated in primary antibody combinations (1:20 for dextran, 1:100 for macrophages, 1:300 for LYVE-1; see figure legends and Table S1 for list of antibodies and suppliers) overnight at 4°C. Primary antibodies were removed with 3 PBS washes and appropriate fluorescently conjugated secondary antibodies (at a 1:300 dilution) were added (45 minutes, room temperature in the dark). Slides were washed, coverslipped with Vectashield liquid mounting media (Vectorlabs) and sealed with nail polish. Sections were imaged using a Zeiss LSM 510 Meta confocal microscope and final, merged-channel images were generated using Image J software.

Iron oxide detection
LNs were subjected to standard immunohistochemistry protocols. Briefly, slides were rehydrated in PBS before endogenous peroxidases were blocked using 3% hydrogen peroxide for 5 minutes. Samples were then subjected to a non-immune block and incubated in blocking buffer (10% Chicken serum in PBS/2% BSA) for 30 minutes, at room temperature. Blocking buffer was removed and replaced with primary antibody solution containing antimouse LYVE-1 (Abcam, 1:300 dilution) for 3 hours at room temperature. Samples were washed 3 times followed by a second non-immune block. Biotinylated anti-rabbit secondary antibody (1:300 dilution) was then added to each section and incubated for 30 minutes at room temperature. Meanwhile, ABC solution (Vectorlabs) was prepared as per manufacturers guidelines. Samples were washed and the pre-prepared ABC solution was added and left on samples for 30 minutes. Following PBS washes, the peroxidase-based substrate solution was prepared and added as per manufactures guidelines (Vectorlabs). At this point samples were closely watched for the development of a brown color resulting from the inclusion of 3, 3'-diaminobenzidine in the substrate solution. When sufficient color developed, the reaction was stopped in miliQ water. Nuclei were counterstained with hematoxylin (Sigma). SPIO was then detected using a commercially available iron stain kit following manufacturers guidelines (Sigma). Sections were imaged using an Olympus BX53 brightfield microscope.

Flow cytometry -tissues
Harvested LNs were gently teased apart with 25G needles and enzymatically dissociated with 1mg/mL collagenase D (Roche). Following incubation at 37°C for 30 minutes the enzymatic reaction was stopped with EDTA. The resulting cell suspension was passed through 70 μm nylon mesh to remove large debris. Samples were transferred to a round bottom 96 well plate where staining was performed. Dead cells were excluded using Live/ dead violet fixable stain (As per manufacturers guidelines; Molecular Probes). Surface antigens were stained with appropriate fluorescently conjugated primary antibodies (all 1:300 dilution), incubating for 30 minutes on ice in the dark. Biotinylated antibodies were detected with conjugated streptavidin (1:300 dilution) for 30 minutes on ice in the dark. In cases where intracellular staining was required i.e. to demonstrate that SPIO was not notably internalized, samples were fixed and permeabilized using a FoxP3 staining kit (eBiosciences) as per manufacturers guidelines, and then incubated for a second time with FITC-conjugated anti-dextran antibody (30 minutes on ice in the dark). Please refer to Table  S1 for list of antibodies and suppliers. Data was then collected on a CyAn™ ADP Analyzer flow cytometer (Beckman Coulter) and analyzed using FlowJo software (Treestar).

Cell culture and nanoparticle incubation
J774.2 (purchased from Culture Collections, Public Health England) and RAW264.7 (kindly provided by Ellen Gokkel) macrophage cell lines were cultured in Dulbecco's modified Eagle's medium containing 4.5 g/L of glucose supplemented with 10% fetal calf serum. Cells were split and incubated with SPIO for 24 hours, at two different concentrations (0.1 mg/mL and 1 mg/mL) and were then thoroughly washed with PBS before subsequent functional analysis. NB; a human dose of SPIO is on average 32 μmol iron/Kg. Here, the 0.1 mg/mL and 1 mg/mL doses were equivalent to 1.79 μmol iron/mL (1.79 mol/Kg -56 times more than in humans) and 17.9 μmol iron/mL (17.9 mol/Kg -556 times more than in humans), respectively. As a control, gold nanoparticles (60 nm, Sigma) were used at a concentration of 4x109 particles/mL. Analysis were performed at 24 hours and 7 days. Nanoparticle uptake was detected by the increase in granularity of the cells, which directly correlates to the increase of the side scatter (SSC-A) of the cells when analyzed by flow cytometry.

Inflammation induction
Macrophages (in presence or absence, or previously incubated with SPIO) were incubated for 24 hours with LPS (bacterial lipopolysaccharide, 100 ng/mL) and INF-gamma (20

Europe PMC Funders Author Manuscripts
Europe PMC Funders Author Manuscripts ng/mL) to induce inflammatory response. Cytokine production and phagocytosis ability was then assessed.

Phagocytosis assay
Large particle phagocytosis ability of macrophages was assessed using 1 μm fluorescent particles (Fluoresbrite YG beads, Polysciences Inc.). Particles were vigorously vortexed and added (1:20 dilution) to a macrophage cell suspension and were then incubated for 20 minutes at 37ºC and 5% CO 2 . Suspensions were washed with PBS and then treated with 0.2% Trypan blue for 3 min and immediately dilute 1:1 in PBS. Two final washes preceded analysis by flow cytometry for particle content. Soluble uptake of antigen was assessed using soluble TRITC (Tetramethylrhodamine)-conjugated ovalbumin (TRITC-OVA, Invitrogen). Macrophages were incubated with 100 μg/mL of TRITC-OVA for 30 minutes. Cells were then washed with ice-cold PBS containing 5% FBS. A final wash with PBS was performed before flow cytometry analysis.

Macrophage polarization analysis -flow cytometry
Flow cytometry was used to analyze the expression of M1 (pro-inflammatory: MHCII, PD-L1, CD86) and M2 markers (anti-inflammatory: Arginase 1, CD163, CD206) on cultured macrophages. Cells were scrapped from the wells following treatments and washed with PBS, and were then incubated with antibodies for the identification of surface markers for 20 minutes at a dilution of 1:200 (primary antibodies) or 1:300 (secondary antibodies). Cell suspensions were then fixed and permeabilized using a fixation/permeabilization kit (eBiosciences) and probed for the presence of intracellular Arg1. Please refer to Table S1 for list of antibodies and suppliers. Data was acquired using a Fortessa (BD) flow cytometer and analyzed using FlowJo.

DQ-OVA assay
DQ-OVA (Invitrogen) was used to determine the ability of macrophage to digest uptaken antigen. Macrophages were incubated with 100 μg/mL of DQ-OVA for 15 minutes at 37ºC, washed three times with ice-cold PBS containing 5% FBS and immediately analyzed by flow cytometry to analyze the presence of the fluorescent GFP product resulting from DQ-OVA digestion.

Proliferation
Cells were seeded in a 12-well plate and proliferation was calculated by following the percent confluency of the well using Incucyte Zoom (Essen BioScience) Live Cell analysis system. Images were acquired every two hours for a total of 90 hours.
As a confirmation methodology cells were seeded in a 12-well plate and counted using the trypan blue exclusion method.

Apoptosis
Cells were assayed in 96-well round bottom plates. After treatment, samples were washed with cold PBS twice and then re-suspended in 100 μL of Annexin V Binding Buffer (cat # 422201, Biolegend). 5 μL of FITC Annexin V (Biolegend) and 5 μL of 7-AAD (cat # 420403, Biolegend) were then added to each well. Samples were incubated in the dark for 15 min at RT (20 °C). 100 μL of Annexin V Binding Buffer was then added to each well and analyzed by flow cytometry. As a positive control, macrophages were incubated with 6 μM of Camptothecin (Acros Organics) for 6 hours prior to Annexin V/7-AAD staining.

Cellular random migration analysis
Macrophage cell lines were seeded in 8-well chamber slides and places in a microscope (Zeiss Axio Observer.Z1 coupled with incubation chamber) where their behavior was followed by acquiring images every 20 minutes for a total of 24 hours. Migration analysis was performed using Volocity software (Perkin Elmer).

Imaging
Bright field images of the different macrophage cell lines were collected using an EVOS (Thermo Fisher) system at day 1 and day 6.

Statistical analysis
Statistical analyses were performed using GraphPad Prism 6 software (GraphPad). For comparisons of two groups, Students t-tests were performed. When comparing 3 or more datasets, One-way ANOVA and appropriate post-hoc tests were performed as described in figure legends.

Supporting Information
Refer to Web version on PubMed Central for supplementary material. Mechanical transport drives rapid SPIO nanoparticle drainage to the lymph node, by a cellindependent mechanism. a) Macroscopic analysis of non-draining and SPIO-draining lymph nodes. b) Representative confocal images of draining lymph nodes 10    used a positive control for apoptosis. e) Quantification of random cell motility after incubation with SPIO and gold nanoparticles. Cells were tracked for 24 hours and parameters of length, track velocity, displacement and meandering index were measured. f) Quantification of markers of macrophage polarization at days 1 and 7 after exposure to SPIO and gold nanoparticles. M1 markers; CD86, MHCII and PD-L1. M2 markers; Arg1, CD163 and CD206. Data presented as Mean ± SD. All graphs are combined data from at least two experiments performed in triplicate. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 (Twoway ANOVA with Tukey post hoc test; in e) One-way ANOVA with Tukey post hoc test).