The clinical application of dendritic cells (DC) as adjuvants in immunotherapies such as the cell-based cancer vaccine continues to gain interest. The overall efficacy of this emerging immunotherapy, however, remains low. Studies suggest the stage of maturation and activation of ex vivo-prepared DC immediately prior to patient administration is critical to subsequent DC migration in vivo, which ultimately affects overall vaccine efficacy. While it is possible to generate mature and activated DC ex vivo using various stimulatory cocktails, in the case of cancer patients, the qualitative and quantitative assessment of which DC stimulatory cocktail works most effectively to enhance subsequent DC migration in vivo is difficult. Thus, a non-invasive imaging modality capable of monitoring the real-time migration of DC in long-term studies is required. In this paper, we address whether cellular magnetic resonance imaging (MRI) is sufficiently sensitive to quantitatively detect differences in the migratory abilities of two different DC preparations: untreated (resting) versus ex vivo matured in a mouse model. In order to distinguish our ex vivo-generated DC of interest from surrounding tissues in magnetic resonance (MR) images, DC were labeled in vitro with the superparamagnetic iron oxide (SPIO) nanoparticle FeREX®. Characterization of DC phenotype and function following addition of a cytokine maturation cocktail and the toll-like receptor ligand CpG, both in the presence and in the absence of SPIO, were also carried out. Conventional histological techniques were used to verify the quantitative data obtained from MR images. This study provides important information relevant to tracking the in vivo migration of ex vivo-prepared and stimulated DC.
Dendritic cells (DC) are one of the most potent antigen-presenting cells (APCs) of the immune system (1), particularly because of their ability to directly prime naive T cells in secondary lymphoid tissues, such as the lymph node (LN) (2). As a result, the use of DC as an adjuvant in immunotherapies such as cell-based cancer vaccines has become an expanding field of research (3–5). Pre-clinical studies in animal models (6–8), as well as early-phase clinical trials (9–11), have provided promising results in regards to the feasibility and safety of the DC-based cancer vaccine. However, the overall efficacy of this emerging immunotherapy, while superior to alternative cancer vaccine strategies (4, 12), remains low and has yet to reach its therapeutic potential (13,14).
Several studies have stressed the importance of the stage of maturation and activation of exvivo-prepared DC immediately prior to administration to a patient (15–18). DC undergoing maturation and activation are characterized by a functional shift from antigen up-take to antigen presentation, requiring an increase in surface expression of antigen presentation molecules (MHC classes I and II), in conjunction with co-stimulatory molecules (CD80 and CD86), to ensure a non-tolerogenic T-cell response. CD40 surface up-regulation is also required for activation of T cells via CD40L ligation (19–24). DC remaining in an immature or semi-mature state can induce tolerance against the presented antigen (25–27). DC exhaustion results when these cells are over-activated before reaching secondary lymphoid tissues (28), often resulting in DC incapable of T-cell activation and premature DC death. As a result, investigations into reagents able to mature and activate DC in vitro effectively without resulting in cell exhaustion have been undertaken. A maturation cocktail introduced by Jonuleit et al. (29) consists of tumor necrosis factor (TNF)-α, IL-1β, IL-6 and prostaglandin E2 (PGE2). This cytokine cocktail (CC) is considered to be a ‘gold standard’ and has been used in several DC vaccine clinical studies (30–32). While this protocol proved effective at maturing and initiating DC activation, the addition of toll-like receptor (TLR) ligands to CCs can enhance the level of DC maturation and activation (33).
DC maturation also induces a shift in chemokine receptor (CCR) expression, resulting in down-regulation of CCRs important for tissue confinement (CCR1 and CCR5) and concurrent up-regulation of surface proteins important for migration toward a LN (CCR7 and CD38) (34–38). Increased surface expression of CD38 by mature DC is important to trigger Ca2+-mediated intracellular signaling cascades required to sensitize surface CCR7 to its chemokine ligands (CCLs) 19 and 21 (34, 39). Therefore, CD38+CCR7+ mature DC migrate efficiently to LNs. It follows, then, that by improving the level of maturation and activation of DC, there will be a corresponding increase in the efficiency with which these DC migrate to LNs (16, 17, 40). As a result, the use of the TLR9 ligand CpG was combined with the standard CC (CpG-CC) in order to mature ex vivo-prepared bone marrow-derived DC (BMDC). This CpG-CC-stimulated DC population was expected to be more efficient at migration toward target LNs in vivo as compared to untreated DC.
In the case of cancer patients, qualitatively and quantitatively assessing which stimulatory cocktail works most effectively to enhance in vivo DC migration is difficult without a non-invasive imaging modality capable of monitoring the real-time migration of DC in long-term studies. We address this issue in a mouse model using cellular magnetic resonance imaging (MRI). MRI is a non-invasive imaging modality used commonly in clinics. It offers exceptional soft tissue contrast and provides three-dimensional images with high resolution and high signal-to-noise ratios. Cellular MRI is an application of this technology, in which cells are labeled with a contrast agent in order to track their fate in vivo. We and others have previously reported the successful application of cellular MRI as a means to monitor mouse (41, 42) and human (43–45) DC migration in vivo. We previously reported that cellular MRI conducted with superparamagnetic iron oxide (SPIO) allows for the semi-quantitative assessment of DC migration (41) and monitoring of DC within the LN (46). Here, we determine if cellular MRI is sufficiently sensitive to quantitatively detect differences in the migratory abilities of DC prepared two different ways: as resting versus ex vivo matured. Furthermore, we determine whether the degree of antigen-loaded DC migration, as assessed by cellular MRI, can be correlated to the magnitude of the ensuing antigen-specific immunological response. In order to distinguish our ex vivo-generated DC of interest from surrounding tissues in magnetic resonance (MR) images, DC were labeled in vitro with FeREX®, a SPIO nanoparticle contrast agent. In order to confirm the presence of SPIO itself has no effect on the maturation and activation of DC induced by CpG-CC, characterization of DC phenotype and function following addition of CpG-CC in the presence and absence of SPIO were carried out. Conventional histological techniques were used to verify the quantitative data obtained from MR images. This study provides important information relevant to tracking the in vivo migration of ex vivo-prepared and stimulated DC. Furthermore, it demonstrates that cellular MRI is capable of detecting differences in the migratory efficiencies of DC matured and activated using different protocols. Ultimately, results of this study can be applied toward addressing important questions regarding DC-based cancer vaccine optimization for use in human clinical trials.
C57BL/6 male mice (6–12 weeks) were obtained from Charles River Laboratories Inc. (Kingston, NY, USA) and housed in pathogen-free conditions in the Robarts Research Institute Barrier Facility (London, Ontario, Canada) until use. All experiments were undertaken in accordance with Animal Care guidelines with the approval from the Animal Use Subcommittee at the University of Western Ontario (London, Ontario, Canada).
RPMI 1640 culture medium (Gibco, Burlington, Ontario, Canada) was supplemented with 10% fetal bovine serum (Hyclone, Logan, UT, USA), 100 U/ml-1 of penicillin and 100 μg ml−1 of streptomycin, 0.3 mg ml−1l-glutamine, 0.01 μg ml−1 MEM non-essential amino acids, 0.1 mM sodium pyruvate, 1 μg ml−1 HEPES buffer solution and 55 μM 2-mercaptoethanol (Invitrogen, Burlington, Ontario, Canada). HBSS was obtained from Invitrogen and BSA was purchased from Sigma–Aldrich (Oakville, Ontario, Canada). Mouse granulocyte macrophage colony-stimulating factor (GM-CSF) and IL-4 were gifts from Dr Peta O’Connell (Robarts Research Institute, London, Ontario, Canada), originally provided by Schering-Plough (Kenilworth, NJ, USA). FeREX® was obtained from BioPAL (Worcester, MA, USA). Anti-mouse CD11c, CD86, CD80, CD40 and I-AB (MHC II) fluorophore-conjugated antibodies were purchased from Becton, Dickinson and Co. (BD, Mississauga, Ontario, Canada). Anti-mouse CD3ε, CD4, CD8α, CD11b, CD19, CD54, CCR7 and IFN-γ fluorophore-conjugated antibodies, biotinylated CD36 and CD38 antibodies and secondary antibodies [streptavidin-PE (SA-PE), SA-allophycocyanin (SA-APC)] were purchased from eBioscience (San Diego, CA, USA). An anti-mouse antibody to the MHC I-SIINFEKL peptide complex was also purchased from eBioscience. Anti-mouse TLR9 was obtained from InvivoGen (San Diego, CA, USA). The H2Kb-SIINFEKL-specific tetramer was purchased from Beckman Coulter (Mississauga, Ontario, Canada). LIVE/DEAD® Fixable Near-IR Dead Cell Stain Kit for use on the FACSAria II was purchased from Invitrogen. Respective isotype controls were purchased from the same companies. Normal goat serum was purchased from Sigma–Aldrich. Recombinant mouse TNF-α was purchased from R&D systems (Burlington, Ontario, Canada), PGE2 was obtained from Sigma–Aldrich, recombinant mouse IL-1β and IL-6 were from PeproTech (Rocky Hill, NJ, USA) and unmethylated CpG (ODN1826) was from InvivoGen.
Generation of murine BMDC
DC used for these experiments were derived from mouse bone marrow precursors. DC were prepared and enriched as previously described (41). Briefly, bone marrow precursors were cultured in complete RPMI 1640 medium for 4 days at 37°C in the presence of GM-CSF and IL-4. Enrichment of DC was performed on day 4 using HistoDenz (13.5% w/v; Sigma–Aldrich) gradient centrifugation.
DC maturation cocktail and SPIO labeling
Following enrichment on day 4, DC were cultured in a maturation cocktail (CpG-CC) and/or labeled with SPIO overnight. The maturation cocktail consisted of TNF-α (25 nM), IL-1β (10 nM), IL-6 (25 nM), PGE2 (10−6 M) and CpG (0.2 μM). For SPIO labeling (FeREX), DC were given 200 μg of Fe/ml. Cells stimulated with CpG-CC but not given SPIO served as the appropriate control. Cells were left to incubate overnight (20 h) at 37°C. Cells left untreated and unlabeled (here after designated as UT) or given SPIO only served as additional controls.
Magnetic separation of SPIO-labeled DC from unlabeled DC
DC were labeled overnight using FeREX SPIO nanoparticles on day 4 as described above. The average labeling efficiency was 76% (46). The total population of unseparated SPIO-labeled DC is here after referred to as SPIOmix, SPIO-labeled DC as SPIO+ DC and SPIO-unlabeled DC as SPIO− DC. If required, SPIO+ DC were magnetically separated from SPIO− DC as previously described (46).
Detection of surface antigens
Cells were collected from overnight cultures on day 5 and prepared for cell surface antibody staining as previously described (41). CCR7 staining was carried out first at 20°C [room temperature (RT)]. All other antibodies were stained at 4°C (on ice). Flow cytometry was performed using a FACSCalibur (BD) and data were acquired with CellQuest Pro (BD) and analyzed using FlowJo software (Tree Star Inc., Ashland, OR, USA). Polychromatic flow cytometry for tetramer analysis was performed using a FASCAria II (BD) and data were acquired using FACSDiva software (BD) and analyzed using FACSDIVA and FlowJo software.
Luminex® cytokine assays
The Luminex® Cytokine Mouse 10-plex panel antibody detection kit (Invitrogen) was used according to the manufacturer’s protocol to analyze the cytokine profiles of supernatants collected from overnight cultures of DC that received CpG-CC ± SPIO. Supernatants collected from UT DC or SPIO+ DC served as controls. Luminex plates were analyzed using the Luminex 100 plate reader and IS 2.3 Software (both from Luminex Corporation, Austin, TX, USA). A minimum of 400 events (beads) was collected for each cytokine per sample.
Labeling DC with PKH
DC were collected from day 5 cultures and washed 1× in cold PBS. DC were stained with 2 x 10−6 M of PKH26 (red) as per manufacturer’s instructions (Sigma–Aldrich) and as previously described (46).
Adoptive transfer of DC
PKH-labeled DC were adoptively transferred via subcutaneous injection into the hind footpads of C57BL/6 mice. Briefly, either 3 × 105 or 1 × 106 PKH-labeled SPIO+ or CpG-CC-SPIO+ DC were administered into the right hind footpads (n = 4). UT DC or CpG-CC DC were used as the respective controls and were administered to the contralateral footpads at the appropriate dose.
MRI of DC migration
Mice were imaged 2 days post-adoptive transfer by MRI using a 1.5 T clinical scanner (GE Medical Systems, Milwaukee, WI, USA) with a custom gradient-coil insert as previously described (41). MR images were analyzed for signal void volume and fractional signal loss (FSL) (41).
Histological analysis and quantification
Following MR scans, popliteal LNs were removed and prepared for conventional histological analysis of DC migration using digital morphometry as previously described (41).
Preparation and administration of SIINFEKL-loaded DC
On day 5, UT, CpG-CC, SPIOmix and CpG-CC-SPIOmix DC were given 10 μg ml−1 of SIINFEKL (MHC I-specific ovalbumin peptide, amino acids 257–265; Sigma–Aldrich) peptide for 2 hours prior to cell collection. DC from each group not given SIINFEKL served as flow cytometric controls. UT DC given hemagglutinin (HA) peptide (Sigma–Aldrich) served as non-specific peptide control cells for flow cytometry. SIINFEKL-SPIOmix DC and CpG-CC-SPIOmix DC were magnetically separated to obtain SIINFEKL-SPIO+ DC and SIINFEKL-CpG-CC-SPIO+ DC.
SIINFEKL-UT, SIINFEKL-SPIO+, SIINFEKL-CpG-CC and SIINFEKL-SPIO+-CpG-CC DC in RT PBS were adoptively transferred into C57BL/6 mice. Each mouse received either SIINFEKL-CpG-CC DC or SIINFEKL-SPIO+-CpG-CC DC to both hind footpads (106 DC per footpad, n = 4). Either 106 SIINFEKL-UT DC or SIINFEKL-SPIO+ DC were administered to both hind footpads of mice (n = 4) to serve as a control.
Preparation of single cell suspensions from LNs
Two weeks following the adoptive transfer of SIINFEKL-UT, SIINFEKL-SPIO+, SIINFEKL-CpG-CC or SIINFEKL-SPIO+-CpG-CC DC DC, mice were euthanized (CO2 inhalation) and both popliteal LNs from each mouse were pooled. LN pairs were digested by using DNase and collagenase type V (Sigma–Aldrich) as previously described (47). LN cells were washed and re-suspended in complete medium for intracellular cytokine staining (ICS) and T-cell stimulation assays.
Tetramer and ICS
Two million LN cells were plated in 96 well plates. Cells were either given media (untreated control), HA (10 μg ml−1) peptide (non-specific peptide control; Sigma–Aldrich), Con A (positive control; 10 μg ml−1, BD) or SIINFEKL peptide (10 μg ml−1). Each treatment was repeated in triplicate (46).
For ICS of IFN-γ positive cells, LN cells were incubated for 2 h at 37°C. Brefeldin A (3 μg ml−1; eBioscience) was then added to cells and plates were incubated for an additional 3 h. Cells were collected and surface antibody staining for CD3ε, CD4 and CD8α was carried out first, followed by intracellular staining for IFN-γ.
For analysis of SIINFEKL-specific T cells, medium was supplemented with recombinant mouse IL-2 (0.2 ng ml−1; PeproTech) and cells were incubated for 2 days at 37°C. LN cells were collected and stained for surface expression of CD3ϵ, CD4, CD8α, CD19 and MHC I-SIINFEKL-specific tetramer as per usual and LIVE/DEAD near-IR staining was carried concurrently.
All data were expressed as means and standard error means. Statistical significance was determined using a repeated measures analysis of variance followed by a post-hoc Tukey’s test. Differences were considered significant if P < 0.05. Correlation between MRI and histological data was determined using a linear rank correlation.
SPIO labeling has no effect on CpG-CC-induced phenotypic and functional maturation and activation of DC in vitro
There has been some controversy regarding the expression of TLR 9 by murine BMDC. Therefore, we used flow cytometry to confirm that our BMDC cultures expressed TLR 9. Following intracellular staining with a mouse TLR 9 antibody consistent with previous reports (48), we found that nearly 100% of UT (untreated and unlabeled) and CpG-CC-treated UT DC groups express TLR 9 (data not shown). Overnight, SPIO labeling had no effect on TLR 9 expression.
Flow cytometry was used to elucidate any phenotypical changes in the cell surface expression of maturation (CD86), functional (CD36, CD11c, CD54, CD80 and MHC II) and activation (CD40) markers of DC following treatment with CpG-CC (Fig. 1A–F). DC left untreated served as a control. The expression of CD36, a functional marker of immature DC to take up apoptotic bodies, was low on mature DC from all groups (Fig. 1D). This was accompanied by significant increases in mean fluorescence intensity (MFI) of surface expressed CD86 (36.1 ± 0.4%), MHC II (47.3 ± 0.7%), CD80 (51.7 ± 0.2%), CD54 (33.9 ± 0.1%) and the activation marker CD40 (56% ± 0.9%) following CpG-CC treatment as compared with UT DC (Fig. 1F). There was a decrease in the MFI of CD36 following CpG-CC treatment. Importantly, SPIO labeling had no significant effect on the CpG-CC-induced maturation and activation of DC (Fig. 1).
Luminex assays were performed in order to determine if CpG-CC treatment affected the cytokine profiles of DC as compared with UT DC (Fig. 1G). CpG-CC treatment of DC resulted in an increase in IL-12 and IFN-γ production as compared with UT DC (P < 0.05). There was also a very small but significant increase in the production of IL-10 by CpG-CC-treated DC. Again, SPIO labeling had no significant effect on the CpG-CC-induced increase in IL-10, IL-12 and IFN-γ (P > 0.05).
CpG-CC treatment induces the expression of surface migration markers by DC in the presence and absence of SPIO
In order to determine if CpG-CC treatment in the presence or absence of SPIO could potentially alter the migratory potential of DC in vivo, we looked at the surface expression of several migration markers of DC using flow cytometry (Fig. 2). The surface expression (MFI) of CD38 on CD11c+CD86− immature DC was significantly lower (28.0 ± 0.4%) following treatment with CpG-CC. The surface expression of CD38 significantly increased by 23.1±0.2% on CD11c+CD86+ mature DC following CpG-CC treatment. The degree of CD11b surface expression (MFI) was not significantly affected by CpG-CC treatment or by SPIO labeling. CCR7 was only expressed by CD11c+CD86+ mature DC, and CCR7 surface expression was not significantly affected by CpG-CC treatment (P > 0.05). CD8α is believed to be expressed by non-migratory DC (49), but this finding is controversial for mouse DC. Other studies claimed that CD11c+CD8α+ define a DC subset with a high migratory capacity in vivo (50, 51). The surface expression of CD8α was negligible on all cell groups (UT, SPIOmix, CpG-CC and SPIOmix-CpG-CC) as compared with the isotype control. Taken together, CpG-CC treatment induced the expression of maturation and activation markers, as well as a migration marker by our mouse DC cultures. This suggests CpG-CC DC should have a higher capacity for migration. Furthermore, SPIO did not alter the expression of these DC surface molecules by CpG-CC-treated DC.
Cellular MRI can detect the difference in migrational ability of different DC populations
Next, we determined if differences in the migration of SPIO+-CpG-CC DC as compared with SPIO+ DC could be detected by cellular MRI. Either 3 × 105 or 1 × 106 SPIO+ DC or SPIO+-CpG-CC DC were adoptively transferred via subcutaneous injection into the right hind footpad of C57BL/6 mice (n = 4). The appropriate dose of either UT DC or CpG-CC DC was injected into the left hind footpads of respective mice to serve as the control. Mice were imaged using cellular MRI 2 days post-adoptive transfer (Fig. 3). MR images were analyzed for LN volume, signal void volume and FSL (Fig. 3B–D). The right poplitieal LNs which received either 1 × 106 CpG-CC DC or SPIO+-CpG-CC DC were significantly larger when compared with those that received either 1 × 106 UT or SPIO+ DC (P < 0.05). The same trend was observed for LNs that received either 3 × 105 CpG-CC DC or SPIO+-CpG-CC DC, but this trend was not significant. The size of the LN also correlated in a dose-dependent manner and was much more evident for the CpG-CC matured DC. Signal void volume increased dose dependently, as did FSL. At both doses, the signal void volumes and FSL were significantly larger for SPIO+-CpG-CC DC when compared with SPIO+ DC by at least 20 and 35%, respectively (P < 0.05). Thus, the data indicate that cellular MRI is sufficiently sensitive to detect differences in DC migration of DC prepared by different protocols.
MRI analysis correlates well with conventional histological analysis of differential migrational capacity of DC populations
In order to verify MR image analysis, conventional histological analysis of DC migration was carried out. This analysis was also used to determine if SPIO labeling had any effect on in vivo migration of CpG-CC DC. There was a dose-dependent increase in DC migration observed (Fig. 4). CpG-CC DC migrated at least 47% better than UT and SPIO+ DC at the 3 × 105 dose. SPIO+-CpG-CC DC migrated 40% better than UT and SPIO+ DC at the 3 × 105 dose, as well (P < 0.05). At the 1 × 106 dose, CpG-CC DC and SPIO+-CpG-CC migrated the most efficiently (42 and 40%, respectively) as compared with the control groups. As we have reported earlier (41, 46), the presence of SPIO reduces the efficiency of DC migration by roughly 20% at the 1 × 106 dose compared with both UT and CpG-CC DC controls. Comparing DC migration data obtained from cellular MRI and conventional histological analysis (Fig. 4D) showed a positive correlation between SPIO+ DC (R2 = 0.70) and SPIO+-CpG-CC DC (R2 = 0.85). Thus, cellular MRI can accurately detect differences in DC content in target LNs.
MHC I loading efficiency using SIINFEKL peptide
UT, SPIO+, CpG-CC and SPIO+-CpG-CC DC were analyzed for their MHC I expression using flow cytometry. It was determined that MHC I was highly expressed on the surface of 100% of CD11c+CD86+ (MFI = 98 ± 12). Although MHC I was present on roughly 65% of CD11c+CD86− immature DC (MFI = 40 ± 3), the level of surface expression was significantly lower than that on mature DC (P < 0.05). This was true for all DC groups tested (UT, SPIO+, CpG-CC and SPIO+-CpG-CC).
Some DC were also pulsed with SIINFEKL peptide 2 h prior to analysis in order to determine MHC I loading efficiency. By employing an antibody that specifically recognizes MHC I-SIINFEKL complexes, we found that 100% of CD11c+CD86+ mature DC from all groups (UT, SPIO+, CpG-CC and SPIO+-CpG-CC) positive for surface expression of the complex, directly reflecting MHC I surface expression. Some UT DC were loaded with an irrelevant peptide, HA in order to verify the specificity of the anti-MHC I-SIINFEKL antibody (Fig. 5D), but non-specific binding was not detected to be different from the isotype control.
Cellular MRI can detect differences in the migration efficiency of different DC populations, which correlates well with resulting immunogenic responses
Two million SIINFEKL-loaded DC (UT, SPIO+, CpG-CC or SPIO+-CpG-CC) were adoptively transferred into C57BL/6 mice via subcutaneous injection into the left and right hind footpads (1 × 106 DC per footpad; n = 4 per treatment). Two days post-adoptive transfer mice that received SIINFEKL-pulsed SPIO+ DC or SPIO+-CpG-CC DC were scanned using cellular MRI. MR images were analyzed as per usual in order to correlate MR-derived image data to observed immunogenic responses (presented below). Two weeks post-adoptive transfer, mice were sacrificed and the popliteal LN cells from each mouse were pooled. LN cells were used for tetramer staining (Fig. 6B) and IFN-γ intracellular staining (Fig. 6D). There was a 49.1% ± 0.1 increase in SIINFEKL-specific CD3+CD8+ cytotoxic T lymphocytes detected in LNs that received either SIINFEKL-CpG-CC or SIINFEKL-SPIO+CpG-CC DC compared to SIINFEKL-UT or SIINFEKL-SPIO+ DC, respectively (P < 0.05). Furthermore, there was a 38.6 ± 0.2% increase in the amount of CD3+CD8+IFN-γ+ T lymphocytes detected from LNs that received either SIINFEKL-CpG-CC or SIINFEKL-SPIO+-CpG-CC DC compared to SIINFEKL-UT or SIINFEKL-SPIO+ DC, respectively (P < 0.05). LN cells from mice that received UT DC stimulated with Con A served as a positive control.
LN cell supernatants were collected from cells used for tetramer staining and analyzed using Luminex® assays to elucidate differences in cytokine levels. The same level of IL-5 was produced by all groups of LN cells, regardless of which DC group was injected (Fig. 6E). It was determined that IL-12 was not produced in significantly high amounts, even following Con A stimulation (Fig. 6G). Although IL-10 was produced by all SIINFEKL-pulsed DC groups (Fig. 6F), the levels were 10× less than that of IFN-γ (Fig. 6H). LN cells that received either SIINFEKL-CpG-CC or SIINFEKL-SPIO+CpG-CC DC produced the highest amounts of IFN-ϒ, nearly 2× that of LN cells which received SIINFEKL-UT or SIINFEKL-SPIO+ DC (P < 0.05).
In order to determine the correlations between MRI data (FSL) and the observed immunogenic outcome, MR images from those mice which received SIINFEKL-pulsed CpG-CC-SPIO+ DC or SPIO+ DC were analyzed for LN volumes, signal void volumes and FSL (Fig. 7A–C). As before, SPIO+-CpG-CC DC FSL and signal void volumes were larger (P < 0.05) as compared with those from SPIO+ DC controls, indicating SPIO+-CpG-CC DC migrated more efficiently. Positive correlations between FSL data and either tetramer or IFN-γ staining were found (R2 = 0.81 or R2 = 0.79, respectively). Ultimately, it was determined that MRI data can be used to detect differences in the degree of in vivo DC migration between different DC preparations. Importantly, these detected differences were shown to correlate well to observed immunological responses.
The targeted migration of clinical grade DC to LNs in cancer patients in vivo is critical for generating effective T-cell-mediated anti-tumor responses. Based on data compiled from recent DC-based cancer vaccine clinical trials, the efficacy of this emerging immunotherapy has not yet reached its full potential. This is in part due to the fact that fewer than 5% of ex vivo-generated DC actually reach target LNs (40). It has been determined that efficient migration to LNs is a hallmark of appropriately matured and activated DC (16, 17). The degree of DC migration also relates quantitatively to the magnitude of the ensuing immune response (52–54). Thus, ex vivo generation of a suitable DC preparation is of significant importance for improving DC-based cancer vaccines. While a number of studies have investigated various pro-inflammatory agents and developed a number of cocktails capable of achieving DC maturation in vitro (29, 55, 56), their relative abilities to improve in vivo migration to LNs in clinical patients is largely unknown. A non-invasive imaging modality with the sensitivity to track and quantify in vivo migration of clinical grade DC in patients would be useful in determining which maturation agents are most beneficial. Potentially, this information could be related qualitatively and quantitatively to the antigen-specific immune responses being elicited by the DC-based vaccine. Therefore, this study set out to determine if cellular MRI is capable of detecting differences between the in vivo migration efficiencies of a mature DC population when compared with a control, resting (un-stimulated) DC population in a mouse model.
Immature mouse DC generated in vitro from bone marrow monocyte precursors express more MHC and co-stimulatory molecules on their surface as compared with native in vivo murine DC due to the presence of GM-CSF and IL-4 in the culture medium (15). Regardless, these DC are not sufficiently mature (neither phenotypically nor functionally) to ensure a proper antigen-specific immunogenic response. In order to generate a suitably mature and activated preparation of DC, incubation with a CC of pro-inflammatory cytokines (IL-1β, IL-6 and TNF-α) plus PGE2 has been used (29). However, these CC-treated DC did not produce significant levels of IL-12-p70, the active form of IL-12. Similarly, DC matured with only TNF-α acted to tolerize CD4+ T cells in vivo (57). Thus, insufficient or inappropriate ex vivo maturation can produce DC that are not optimal for use in cell-based immunotherapies (16,27).
DC recognize evolutionarily conserved pathogen-associated molecular patterns (PAMPs). PAMPs are recognized by DC by pattern recognition receptors (PRRs) or TLRs. Stimulation of DC using PAMPs induces DC maturation along with appropriate Th 1 type cytokine production. Recently, Napolitani et al. (58) demonstrated that TLR ligation leads to the induction of high amounts of IL-12, leading to a Th 1 polarized helper T-cell response. It has also been demonstrated that the use of TLR ligands in combination with pro-inflammatory cytokines is able to produce a DC population with a more mature and active phenotype (33). Taken together, data suggest the use of the gold standard CC in conjunction with TLR ligation might optimally induce DC maturation that supports efficient migration to LNs and Th 1 type responses required for an effective DC-based vaccine.
The combination of CpG with CC (CpG-CC) stimulation resulted in the up-regulation of maturation and activation markers MHC II, CD86, CD80, CD54 and CD40 but an insignificant increase of CCR7, a key mediator of mature DC migration. It has been determined recently that while surface expression of CCR7 is required for DC migration to CCLs 19 and 21 (CCL19 and CCL21), it is not sufficient (59, 60). A study investigating the role of PGE2 suggested that CCR7 pathways require sensitization to their ligands by activation of Ca2+ channels, mediated by the ecto-enzyme CD38 (59). CpG-CC stimulation did result in an up-regulation of CD38 surface expression by our mouse BMDC cultures. The surface expression of the integrin CD11b, which has been shown to play an important role in DC adhesion to reticuloendothelial cells and extracellular matrix molecules (61), remained largely unaffected by CpG-CC stimulation. We also did not observe any induction of CD8α expression, which is believed to be expressed by non-migratory DC (49). Importantly, these experiments demonstrate that simultaneous labeling with SPIO had no affect on CpG-CC stimulation of our DC in vitro.
We did not observe any effect of SPIO on CpG-CC-induced DC cytokine secretion. It is accepted that Th1 type responses play important roles in the generation of cellular immunity and are most useful in terms of cancer immunotherapy (1,17,40). Certain key cytokine mediators play important roles in the control of the Th1/Th2 balance including IL-2, IL-12, IFN-γ, IL-4 and IL-10. IL-12 and IL-10 production play an important role in determining the balance between immunogenic and tolerogenic responses (62–64). CpG-CC stimulation increased the production of IL-12 while having a relatively very small but significant effect on IL-10 secretion by DC in vitro. SPIO did not alter this response, suggesting SPIO+-CpG-CC DC should be capable of eliciting a Th 1-type response in vivo.
Previous studies have indicated that MRI is sensitive enough to detect low numbers of cells in vivo (41,65), which was confirmed again in this study. There was a strong correlation between quantitative MRI analysis and conventional histological digital morphometric analysis of the differences in DC migration exhibited between SPIO+ and CpG-CC-SPIO+ DC. This demonstrates that cellular MRI had sufficient sensitivity to detect the increased migration of the SPIO+-CpG-CC DC as compared with that achieved with resting SPIO+ DC. Notably, as demonstrated by histological analysis, all DC populations tested migrated to central areas of LNs, thereby maximizing their potential to activate naive T cells.
For the first time, we demonstrate that it is possible to correlate cellular MRI data of DC migration to immunological outcomes. Ex vivo T-cell responses were assessed using either an antigen-specific tetramer or an intracellular IFN-γ staining using LN cells from mice that were imaged using MRI. As suggested by another report (54), we expected the degree of enhanced migration of SPIO+-CpG-CC DC should correspond to an enhanced T-cell response as compared to control SPIO+ DC.
The use of SPIO as a cellular MRI contrast agent was well established in a number of proof-of-principle pre-clinical and clinical studies (41–43, 46). These studies used either the United States Food and Drug Administration (FDA)-approved Feridex® or the related European version, Endorem®. Unfortunately, Feridex was withdrawn from the market in November 2008 (66), and while Endorem is presently available in Europe for clinical use, it is not approved for use in North America. This study used FeREX, an FDA unapproved SPIO nanoparticle of similar physical and chemical composition to Feridex. Feraheme® has been investigated as a potential replacement for Feridex® as a cell tracking contrast agent. An FDA-approved ultra-small SPIO developed as an iron replacement agent. Feraheme was designed for rapid uptake and degradation by the liver (67). Initial indications suggest that Feraheme may not be not useful for tracking DC due to its shorter half-life in the cell, lower cell up-take and its lower iron content per particle [unpublished data and (68)]. As of yet, there is an apparent need for a new good manufacturing practice-grade FDA-approved iron-based contrast agent with sufficient sensitivity required to track DC or other cells used in cell-based therapies.
Recently, 19F-based contrast agents have received attention for their use in DC cell tracking studies in pre-clinical animal models (44, 69, 70). The advantage of 19F agents over SPIO nanoparticles is that 19F generates positive signal that is easily quantifiable over a long linear signal range, while SPIO generate a negative contrast signal void with a short quantifiable linear signal range (70). However, the detection sensitivity provided by SPIO nanparticles is superior in comparison with 19F agents—at least 10-fold more sensitive (71,72). In vivo detection of a single SPIO-labeled cell has been demonstrated in images acquired on low field strength 1.5T clinical scanners (65). When compared with SPIO contrast agents, cell tracking with 19F agents requires higher field strength MR scanners. It remains to be determined whether 19F-based contrast agents prove clinically useful in humans.
In summary, it was observed that SPIO did not negatively affect CpG-CC-induced up-regulation of key co-stimulatory, migratory and activation molecules on mouse BMDC. Results from this investigation provide evidence that cellular MRI has the sensitivity to track DC in vivo and to quantitatively determine differences in migration of different DC preparations. Furthermore, MRI data can be correlated to immunological outcomes. This investigation provides the proof-of-principle concept that cellular MRI can be used as a modality to optimize DC cultures for improved clinical outcomes.
S.D. was supported by the Frederick Banting and Charles Best Canada Graduate Scholarship Master’s Award (Canadian Institute for Health Research); Translational Breast Cancer Research Trainee Studentship (London Health Sciences Centre); This research was supported by the Ontario Institute for Cancer Research; and the Terry Fox Foundation for Cancer Research.