Synthetic amorphous magnesium-substituted calcium phosphate (sAMCP) nanomineral particles were prepared as previously described (20). Briefly, for the preparation of particles incorporating PGN (Staphylococcus aureus, Fluka), PGN was added to phosphate (PO₄) solution [containing 39 mM (NH₄)HPO₄ (Sigma-Aldrich) in 0.15 M TRIS buffer, adjusted to pH 8 with hydrochloric acid], prior to mixing in equal parts with calcium–magnesium–BSA solution (BSA; Sigma-Aldrich Company Ltd., Dorset, UK) dissolved in a Ca/Mg solution [containing 35 mM CaCl₂ (Sigma-Aldrich), 7.2 mM MgCl₂ (Sigma-Aldrich) in 0.15 M Tris buffer (Sigma-Aldrich)] at 1 mg/mL for a final PGN concentration of 50 μg/mL mixed solution. For particles incorporating protein purified derivative (PPD) of tuberculin (Statens Serum Institute, Denmark), PPD was added to the phosphate solution prior to mixing with the calcium–magnesium–BSA solution for a final PPD concentration of 100 μg/mL, which was then incubated with gentle rotation at room temperature for 1 h to allow particle formation. Particles were precipitated by centrifugation (1,500 rpm for 5 min), washed in pH 10 water, followed by washing in tissue culture grade water before re-suspending in tissue culture media at half of the original particle solution volume.

Size distribution of amorphous calcium phosphate particles was examined using a Nanosight NS500 instrument and NTA 2.1 software (NanoSight Ltd.). Freshly synthesized particles were diluted 1:10 to an optimal particle concentration of between 10⁷ and 10⁹ particles/mL immediately prior to loading into the viewing chamber. Camera levels and focus were adjusted for optimal particle visualization and tracking. A capture duration of 90 s was used to record each video for NTA analysis. Mode values were obtained through area-under-curve analysis in GraphPad Prism 6.03 while average size and percentiles were determined in Microsoft Excel. For the latter, data were transformed into cumulative percentage and exact D10, D50, and D90 values calculated with linear interpolation using the two data points above and below the 10th, 50th, and 90th percentile, respectively.

Following synthesis, empty particles or particles containing BSA, PPD, and PGN (alone or in conjunction) were centrifuged at 1,500 rpm (5 min) and supernatants were collected. After further washing twice, in pH 10 water, particles were dissolved in 100 mM citric acid buffer (pH 3) to release the organic material. Total protein content was measured using the Bradford protein assay (as per manufacturer’s protocol, Bio-Rad Laboratories, UK) while PGN matter was quantified using both a colorimetric assay and a modified Periodic Acid Schiff (PAS) assay as described (21).

Anonymized, snap frozen human ileal tissue specimens, containing lymphoid patches, were purchased from a commercial tissue bank (Tissue Solutions, UK) and held for use by the authors under HTA license agreement 12383. Samples were from the normal resection margins of three patients with tumors (Table 1). Tissue sections were cryo-sectioned at 14 μm thickness, collected on SuperFrost^® slides (Thermo Scientific, USA) and allowed to air dry for 30 min at room temperature.

Tissue sections were fixed in formaldehyde (Sigma-Aldrich, UK) and blocked (10% v/v goat serum, 2% w/v BSA, 0.3 M Glycine). Sections were then incubated with the primary antibody against human IL-10 (ab34843; Abcam, UK), followed by incubation with a secondary antibody (Alexa Fluor^® 568 Goat Anti-rabbit IgG [H + L], Invitrogen, UK) and stained for the nanomineral using 0.01 M calcein (Sigma-Aldrich, UK). Finally, the nuclei were counterstained with Hoescht 33342 (H1399, Invitrogen, UK). Sections were imaged with a Leica DMIRE2 microscope (Leica Microsystems, Germany) at 405, 488, or 568 nm, fitted with diode, Ar/ArKr, and HeNe lasers, using a 63×, 1.2 NA water objective lens. Data were recorded using the Leica Confocal Software (v2.61) and images processed using Imaris version 8.0.0 (Bitplane, UK). Data were collected as 8-bit gray scale images and assigned appropriate colors subsequently.

The human blood cells obtained for this study were single leukocyte cones purchased from the National Blood Service (UK), peripheral blood mononuclear cells (PBMC) were isolated using density gradient centrifugation. Following isolation, PBMC populations were sometimes then stored (in the short term) at −80°C until required. Cell stimulation and incubations were carried out at 10⁶ cells/mL in RPMI 1640 Media (Sigma-Aldrich) containing 10% fetal calf serum (FCS) (PAA Laboratories), 100 U/mL penicillin, 100 μg/mL streptomycin, and 2 mmol/L l-glutamine (all Sigma-Aldrich), and incubated at 37°C, with 5% CO₂. Soluble stimulants were added at concentrations of 5 μg/mL PGN (S. aureus, Fluka) and/or 10 μg/mL PPD of tuberculin (Statens Serum Institute #2391, Denmark). Controls included 50 μg/mL BSA in PBS (low endotoxin, Sigma-Aldrich) and 1 μg/mL staphylococcal enterotoxin B (positive T cell assay control; Sigma-Aldrich, data not shown). All particulate stimulations were performed by re-suspending 100 μL of freshly synthesized particles in cell culture media containing 10⁶ cells/mL. For cell incubations above 3 h, cells were washed with RPMI and replenished with fresh media (only) and returned to 37°C for the remaining period of incubation. After incubation, cells were centrifuged at 1,500 rpm for 5 min to pellet cells and the supernatant was removed and stored frozen at −80°C until multi-analyte analysis performed by Myriad RBM (Austin, TX, USA), or ELISA according to manufacturers’ instructions (R&D Systems Europe Ltd., Abingdon, UK).

For surface marker staining, at the end of the incubation period, cells were centrifuged at 1,500 rpm for 5 min to pellet cells and the supernatant removed. The cell pellet was stained with either anti-PD-L1 FITC, PD-L2 PE, and CD14 PerCP Cy5.5, for costimulatory panels or human leukocyte antigen (HLA)-DR FITC and CD14 PerCP Cy5.5 (all BD Biosciences, UK) according to manufacturers’ instructions for 20–40 min on ice in the dark. Cells were then washed with ice cold PBS 1% BSA, re-suspended in a small volume of PBS containing 1% PFA and placed on ice in the dark until acquisition. Single stain compensation tubes for each stain used, as well as an unstained tube, were also prepared from cell samples at this time in order to compensate for spectral overlap. Cells were filtered immediately prior to acquisition on a Cyan-ADP flow cytometer using Summit V4.3.02 software for acquisition and Summit V4.3 for analysis (Beckman Coulter, UK), acquiring a minimum of 250,000 events per sample.

For flow cytometric analysis of particle internalization, particles were prepared as described with the addition of 1 μL of 10 mg/mL Calcein stock solution (Sigma-Aldrich, UK) added to the phosphate solution prior to mixing for each milliliter of mixed solution. Mixed solutions were protected from light thereon and further incubated with rotation at room temperature for 1 h followed by the normal centrifugation and washing once in tissue culture grade water. Then, 100 μL of fluorescently tagged particles (~1.2 × 10⁸ to 1.6 × 10⁸ particles/mL) were added to 10⁶ cells and incubated for 3 h, before washing and re-suspending the cells in ~200 μL ice cold PBS 1% BSA. Cells were then stained for 20 min on ice in the dark for CD14 PerCP Cy5.5 (BD Biosciences, UK) and CD3 Vio-Green (Miltenyi Biotec, UK).

Proliferation assays were carried out as previously described (23). Prior to stimulation, PBMC were stained at 2 × 10⁶/mL with 0.1 μM CFSE (Sigma-Aldrich) diluted in sterile PBS, for 7 min at 37°C in the dark and then washed three times with RPMI 1640 containing 20% FCS. From then onward cells were protected from light. CFSE-stained PBMC (1 × 10⁶ cells/mL) were transferred into FACS tubes and incubated with relevant stimulants as described. On day 5, cells were washed and stained for the surface markers CD3 PE Cy5 and CD4 PE (BD Biosciences, UK) according to manufacturers’ instructions; 500,000 events were acquired for each sample as described previously.

For Imagestream analysis of particle internalization, particles were synthesized to incorporate calcein, PPD, and PGN as previously described. A 100 μL of fluorescently tagged particles were added to 10⁶ cells/mL and incubated for 3 h, before washing and re-suspending cells in ice cold PBS at 10⁷ cells in 100 μL. Cells were then stained for 20 min on ice in the dark for the surface marker CD14 PerCP Cy5.5 (BD Biosciences, UK), washed, re-suspended in a small volume of PBS, and filtered before acquisition on an ImagestreamX using INSPIRE V4.1.501.0 software (Amnis-Merck Millipore, USA), acquiring a minimum of 50,000 events per sample. A stained particle control was incubated with cells and acquired in addition to unstained particle/cell negative controls and single surface marker controls required for the generation of a compensation matrix and analysis using IDEAS V6.0 software (Amnis-Merck Millipore, USA).

Statistical evaluation where multiple comparisons were performed were assessed by one way ANOVA and post hoc analysis using Tukey’s honestly significant difference method with significance taken as P = < 0.05 (particle uptake, induction of surface receptor expression, cytokine production, and T cell proliferation). Only statistically relevant differences are discussed in the text. Experiments comparing two datasets were performed using paired, Student’s t-test. P = < 0.05 was considered statistically significant.

To examine the influence of synthetic AMCP (sAMCP) antigen carriage on CD4⁺ antigen-specific T cell responses, sAMCP particles were synthesized and characterized as recently reported (20), but in the presence of PPD of tuberculin from Mycobacterium tuberculosis. Resulting particle size distribution, assessed using nanoparticle tracking analysis, revealed a hydrodynamic size (D10–D90) ranging from 67 to 320 nm, a D50 of ~185 nm and peak sizes of about 130 and 210 nm (Figure 1A; Table 2). In some experiments, sAMCP particles were also synthesized to additionally incorporate the bacterial component PGN together with PPD and nanoparticle tracking analysis confirmed that this did not significantly alter the particle size range (i.e., D10–D90: 55–350 nm and D50 of 174 nm, Figure 1B; Table 2, peak sizes of 140 and 270 nm). Thus, overall, sAMCP bearing PPD or PPD together with PGN combined within the same particles demonstrated a similar size range to the naturally occurring endogenous nanomineral (2) (i.e., 75–150 nm) with a small proportion reaching larger agglomerates of ~350 nm, as recently detailed for sAMCP with ovalbumin incorporation (20). The extent of incorporation of PPD antigen was quantified using the PAS assay while trapping of RBB-dyed PGN was assessed directly using colorimetry. With this protocol, sAMCP incorporated around 50% of the total PPD and 90% of the total PGN added during synthesis (Figure 1C). Taken together these data demonstrated high particle loading of our reporter protein antigen (PPD) while maintaining intestinal particle characteristics and size range. In subsequent stimulation assays, particle concentrations were adjusted such that PBMC were exposed to final concentration of ~10 μg/mL PPD and PGN.

In vivo, endogenously formed AMCP nanoparticles deliver luminal antigen to gut tissue APCs (2). We first sought to examine whether AMCP delivery of antigen influences antigen-specific T cell responses, compared to APC uptake of the same antigen in soluble form. For this we adapted methods described by Singh and Booth, using the MHC class II and costimulation dependant CD4⁺ T cell recall antigen PPD to examine antigen-specific T cell responses (24). In brief, interferon-gamma (IFN-γ) production and proliferation of CD4⁺ T cells within PBMC of PPD-responding donors was measured at 5 days of culture, after a 3 h stimulation of the PBMC with soluble or sAMCP-incorporated PPD.

Synthetic AMCP carriage of PPD, failed to induce either robust secretion of IFN-γ or CD4⁺ T cell proliferation in PBMC of PPD-responding donors, despite matched loading of nanominerals with the soluble PPD dose. By contrast, exposure to soluble PPD induced both a strong IFN-γ response and proliferation of CD4⁺ T cells (Figures 2A–C). The reduction of T cell responses, when sAMCP-incorporated antigen was taken up by APCs, was not restricted to PPD as the same result was observed with Tetanus Toxoid (Figure S1 in Supplementary Material). These findings were unexpected as we have previously shown that sAMCP is an effective delivery vehicle for APCs, is inert itself, releases its cargo intracellularly, and does not attenuate responsiveness to PGN at the gene level (20). We therefore investigated the uptake of, and CD4⁺ T cell responsiveness, toward sAMCP particles containing both PPD and PGN.

First, we examined the efficiency of uptake of PPD + PGN-loaded sAMCP by APCs present in PBMC. Synthetic AMCP particles were synthesized as described in the previous sections, but in the presence of the fluorescent marker calcein (20), and loaded with PPD and/or PGN. Control sAMCP particles were loaded with BSA only. All particles were incubated with PBMC for 3 h. Using flow cytometry and flow imaging techniques internalization of sAMCP was assessed in both CD3⁺ T lymphocytes and CD14⁺ monocytes. Calcein⁺ particle staining, representing AMCP uptake, corresponded almost exclusively with cells of the monocytic lineage. Almost 60% of CD14⁺ monocyte/macrophage cells were positive for calcein after 3 h sAMCP exposure, compared to less than 1% of the CD3⁺ lymphocyte population (Figure 3). Flow imaging confirmed that sAMCP–PPD–PGN particles were taken up more readily than sAMCP control particles (sAMCP–BSA), and that the particles were not just associated with monocytes but were internalized, with 51.1 ± 10.5% SEM defined as highly internalized, visualized residing well within the CD14-defined cell membrane boundary (Figure 4).

As well as being pro-inflammatory, under certain circumstances PGN is known to play important roles in countering inflammatory responses and inducing tolerogenic responses through the secretion of IL-10 and the expression of PD-L1 by APCs (10, 13, 25–28). Hence, we next studied the effects of sAMCP-delivered PPD/PGN on primary PBMC cultures in terms of (i) the induction of the cell surface regulatory coreceptors PD-L1 and PD-L2, (ii) IL-10 secretion, and (iii) modulation of the HLA which is associated with MHC class II presentation of exogenously derived antigen. The delivery of PPD in AMCP form (but not in suspension) increased PD-L1 expression by CD14⁺ APCs. As anticipated (and in agreement with previous studies where PGN was present in suspension), the addition of PGN also stimulated the expression of PD-L1, but not PD-L2, by the APCs and additionally stimulated IL-10 secretion (Figures 5A,B). Furthermore, cell surface expression of the MHC class II antigen presenting molecule, HLA-DR, was significantly increased on APCs in response to PPD by 24 h only when PGN was absent (Figure 5C).

We therefore considered the mechanism by which intracellular delivery to APCs of PGN plus PPD antigen could influence the proliferation of antigen-responding T cells within PBMC derived from a further six healthy, PPD-responding, donors. Antigen-specific CD4⁺ T cell proliferation assays were performed as before (Figure 2) but with the addition of blocking antibodies against either PD-L1 or the IL-10 receptor (IL-10r). Consistent with our previous results, delivery of PPD in sAMCP reduced CD4⁺ T cell proliferation, and this inhibition was further enhanced when PGN was present (Figure 6). Blocking the IL-10r and hence inhibiting IL-10 activity restored T cell proliferation, whereas the neutralization of PD-L1 had no such effect (Figure 6). Taken together our data show that multiple checkpoints appear to inhibit T cell proliferation in response to sAMCP-delivered antigen. First, the nanomineral itself does not permit efficient protein presentation for T cell proliferation in comparison to delivery of free protein. Second, the concomitant carriage of PGN, which occurs in vivo, reduces the available HLA-DR on the surface of the APC thereby limiting antigen presentation via the effect of IL-10 production. Finally, antigen that is presented occurs in the context of immuno-inhibitory PD-L1 (Figure 5).

These studies indicated that the secretion of IL-10 provided the most striking attenuation of antigen-specific CD4 T cell proliferation in response to nanomineral carried PPD–PGN. To verify the physiological relevance of these observations, we therefore sought to confirm that AMCP nanomineral-bearing cells in the Peyer’s patch sub-epithelial dome (SED) exist in an IL-10 high environment. Samples of human tissue, from three patients, containing Peyer’s patches were collected from the uninvolved margins of surgically resected intestinal specimens. AMCP⁺ cells of the SED region were detected by calcein staining and IL-10 protein was detected by specific antibody-based staining. Consistent with previous reports, all specimens showed clear calcein staining in the SED representing the presence of AMCP (29). In addition, IL-10 staining was strongly positive in the same area compared to the rest of the patch (Figure 7). Thus, our findings imply that the failure to induce antigen-specific CD4 T cell proliferation, observed in vitro in response to antigen and PGN codelivered to APCs within sAMCP nanomineral, is of physiological relevance in vivo as the key attenuating molecule (IL-10) appears abundantly present in close proximity to AMCP.