Neutrophil vaccination dynamics and their capacity to mediate B-cell help in Rhesus Macaques

Abstract

Neutrophils are the most abundant leukocyte and play a critical role in the initial response to an antigen. Recently, their ability to contribute to adaptive immunity has been highlighted. We evaluated the ability of neutrophils from blood to contribute to the adaptive immune response in a pre-clinical Rhesus macaque SIV vaccine trial. Replication-competent adenovirus-SIV recombinants induced neutrophil activation, B-cell help markers, and enhanced ability to generate reactive oxygen species. Boosting with SIV vaccines (adjuvant together with ALVAC or DNA plus envelope protein) elicited significant neutrophil responses. Serum cytokine and chemokine levels induced correlated with the frequency of neutrophil subsets expressing IL-21, myeloperoxidase, and CD64. Post-SIV infection, neutrophils exhibited dysfunction, both phenotypically and functionally. B-cells from protected and infected macaques co-cultured with autologous polymorphonuclear cells (PMNs), consisting primarily of neutrophils, were activated, underwent class switching, and produced antibodies. This B-cell help was not aided by addition of IL-10 and was largely contact-dependent. Numerous genes associated with inflammation, antibody production, and chemotaxis were upregulated in the co-cultured B-cells. We conclude that immune stimulation by vaccination or antigenic exposure imparts a greater ability of neutrophils to contribute to the adaptive immune response. Harnessing this granulocytic response has the potential to improve vaccine efficacy.

Introduction:

Neutrophils are the most prevalent leukocyte and exert considerable influence on the innate immune response, with increasing evidence that they also contribute substantially to adaptive immunity (1). Their innate functionality as granulocytes entails the release of a vast array of cytokines and chemokines (2). They are stimulated by various chemoattractants and subsequently traffic to sites of inflammation, where they can actively kill invading pathogens via phagocytosis, degranulation, or by releasing neutrophil extracellular traps (NETs) (3). They contribute to adaptive immunity through immune cell crosstalk that can be both immunostimulatory and immunoregulatory, as well as by aiding in the resolution of inflammation (4). Recently, it was clearly demonstrated that both human and rhesus macaque neutrophils can act as APCs, presenting antigen in vitroor vaccine antigen ex vivoto CD4⁺T-cells (5, 6). Although neutrophils are not often studied in the context of HIV and SIV infection (7), the diversity of their functions, and the breadth of their effects on immune responses intimate that they could play a vital role in both HIV/SIV vaccination and viral pathogenesis.

Neutrophils exhibit a complex response to HIV. They are activated by HIV-1 (8), and even by HIV single stranded RNA alone (9). In fact, neutrophil expression of CD64 (FcγRI) has been proposed as a marker of systemic inflammation following HIV infection (10). During HIV infection, there is a generally observable dysregulation of various granulocyte functions (7). Despite this dysfunction, neutrophils can still act directly against HIV via NETs (11), generation of reactive oxygen species (ROS) (12, 13), and phagocytosis (14). This effector functionality targeted against HIV, as well as the dysfunction caused by HIV infection, are significant aspects of the immunological response of neutrophils to HIV. Both should be understood in the context of HIV vaccine development, particularly as they relate to one of the main goals of vaccination: the elicitation of protective HIV antibodies.

Vaccine induction of antibody is directly dependent on how B-cells are affected by the vaccine. Recently there has been widespread interest in the ability of neutrophils to mediate B-cell help and contribute to immunoglobulin production. Neutrophils may contribute to antibody induction by collecting antigens at sites of inflammation (15). They are also sources of BAFF and APRIL (16–18), factors which promote survival and differentiation of B-cells. In humans, it has been demonstrated that splenic neutrophils induce class switching and antibody production by marginal zone B-cells through a mechanism involving IL-21, BAFF, and APRIL (17). While circulating neutrophils appeared unable to contribute significantly to B-cell help, when exposed to sinusoidal endothelial cells which expressed IL-10, they gained this ability. Splenic B-cell helper neutrophils have also been demonstrated in vivoin mice, activating B-cells via pentraxin 3 (19). This ability of neutrophils to mediate B-cell help warrants further experimentation, particularly in the context of mucosal and systemic immune stimulation, as occurs during vaccination and HIV/SIV infection.

This study explores neutrophil responses and their influence on adaptive immunity over the course of a pre-clinical SIV vaccine study in rhesus macaques extending from pre-vaccination, through heterologous prime-boost immunizations, SIV challenge exposures, and subsequent acute and chronic infection or protection. We report that the neutrophil response to vaccination consists of both phenotypic changes and alterations in their functional ability to respond to antigen. Their response to infection is largely in accordance with previous experimental observations regarding neutrophil dysfunction. Importantly we show that when PMNs from blood are co-cultured with autologous B-cell enriched PBMCs, they elicit B-cell help. The B-cells exhibit signs of class switching and blasting, and also produce antibodies, when co-cultured with PMNs. These data suggest that immune stimulation of neutrophils via vaccination or other antigenic stimuli can contribute significantly to the adaptive immune response against that same immune stimulation.

Methods:

Animals, immunization and challenge

Sixty Indian rhesus macaques (Macaca mulatta) aged 3 to 4 years and negative for SIV, SRV, and STLV were used in this study (Musich et al., in preparation) as outlined in Supplemental Figure 1. Macaques were primed at weeks 0 (intranasally and orally) and 12 (intratracheally) with replication-competent adenovirus type 5 host-range mutant (Ad5hr) recombinants (5 X 10⁸pfu/dose/route) separately encoding SIV_M766gp120-TM and SIV₂₃₉gagor empty Ad5hr vector_.as previously described (20). Subsequently at weeks 24 and 36 macaques in the ALVAC/Env group were boosted with ALVAC-SIV_M766gag/pro/gp120-TM (10⁸pfu) and 200 μg each of gD-SIV_M766gp120 and gD-SIV_CG7Vgp120, both in alum hydroxide, intramuscularly (IM). Macaques in the DNA&Env group were boosted with DNA encoding SIV_M766gp120-TM (1 mg), SIV₂₃₉gag (1 mg), and macaque IL-12 (0.2 mg) IM, followed immediately by electroporation and administration of the same SIV gD-gp120 proteins in alum phosphate at the same sites. Control animals received alum hydroxide or alum phosphate only. At week 42, all macaques were challenged intrarectally using a repeated low dose of SIV_mac251(1:500 dilution; 120 TCID₅₀), a challenge stock developed by Dr. Ronald Desrosiers and provided by Dr. Nancy Miller, Division of AIDS, NIAID. Up to 15 challenges were continued weekly until the onset of infection determined by a plasma viral load of ≥50 SIV RNA copies/ml as assessed by the NASBA method (21, 22). Macaques were monitored at least 40 weeks after infection or until euthanasia criteria were met. 6 naive rhesus macaques were also used where indicated.

Tissue Preparation

Lymph node biopsies were collected, minced, and passed through a 40-μm cell strainer before contaminating RBCs were lysed. Cells were washed and resuspended in R10 medium (RPMI 1640 with 10% FBS) containing 2mM L-glutamine, 1% antibiotic–antimycotic (Gibco). Cells were stained fresh.

Rectal pinch biopsies were processed as previously described (23). Briefly, rectal pinches were rinsed with pre-warmed RPMI1640 (Invitrogen) containing 2×antibiotic–antimycotic solution, 2-mM L-glutamine (Invitrogen) and 2 mg/ml Collagenase (Sigma–Aldrich). Prior to incubation (25 min at 37°C) the pinches were minced using a scalpel and a 19G needle, transferred in 10 ml of the same media to a 50 ml tube and pulse vortexed every 5 min. The digested tissue was passed 5 times through a blunt end cannula. The liberated cells and tissue debris were passed through a 70 μm cell strainer, and the cells were washed in R10 containing 2×antibiotic–antimycotic solution and 2 mM L-glutamine prior to staining.

Whole Blood Staining

Anti-human fluorochrome-conjugated mAbs known to cross-react with Rhesus macaque antigens were used in this study, including PE anti-CD182 (6C6), PE-Cy5 anti-IgG (G18–145), PE-CF594 anti-CD32 (FL18.26), PE-CF594 anti-CD64 (10.1), AX700 anti-Ki67 (B56), BV605 anti-CD162 (KPL-1), BV605 anti-CD138 (MI15), v450 anti-IgM (G20–127), BV711 anti-CD197 (150503), BV786 anti-CD45 (D058–1283), BV786 anti-CD14 (M5E2), BUV395 anti-CD184 (12G5), BUV737 anti-CD64 (10.1), and BUV805 anti-CD14 (M5E2) (all from BD Biosciences, San Jose, CA); PerCP-eFluor710 anti-CD274 (MIH1), PerCP-eFluor710 anti-CD27 (O323), APC-Cy7 anti-CD11b (ICRF44), efluor450 anti-MPO (MPO455–8E6), and eVolve655 anti-CD20 (2H7) (all from eBioscience, San Diego, CA); FITC anti-CD2 (RPA-2.10), PE anti-CD257 (T7–241), PE-Cy7 anti-CD80 (2D10), APC anti-HLA-DR (L243), Ax647 anti-IL-21 (3A3-N2), Ax700 anti-TNFα (MAb11), and Ax750 anti-CD69 (FN50) (all from BioLegend, San Diego, CA); Ax488 anti-AID (polyclonal Rab) (Bioss); TxRed anti-IgD (Goat IgG) (Southern Biotech); PE-Cy7 anti-CD66abce (TET2) (Miltenyi); and PE anti-CD38 (OKT10) (NIH Nonhuman Primate Reagent Resource, Boston, MA). MPO, IL-21, TNFα, and Ki-67 were all stained intracellularly. The Aqua LIVE/DEAD viability dye (Invitrogen, Carlsbad, CA) was used to exclude dead cells. 125μL of EDTA whole blood was stained with live/dead aqua for 15 minutes at room temperature (RT). Surface staining antibodies were then added, and blood was incubated for 25min at RT. 1X BD Facs Lyse was then added and incubated with cells for 5 minutes at RT. Cells were washed with 2% FBS in PBS prior to resuspending in BD Fix-Perm and incubating for 15min at RT. Cells were then washed with BD Permwash, and intracellular staining Abs were added to cells and incubated for 25min at RT. Cells were then washed with PBS and fixed with 2% Formaldehyde prior to acquiring staining data using either a BD LSRII or a BD Fortessa.

PMN/Neutrophil Isolation

Following PBMC isolation (see below), all granulocytes were obtained from the erythrocyte/granulocyte layer after Ficoll-Paque density gradient centrifugation and diluted in PBS. The granulocytes were then mixed 1:1 with sterile 6% Dextran in PBS solution by shaking and placed at RT for 1hr to allow sedimentation of RBCs. The top layer containing granulocytes was then washed with PBS, and remaining RBCs were lysed with RBC lysis buffer (155mM NH₄Cl, 12mM NaHCO₃, and 0.1 mM EDTA), washed with PBS, and resuspended in R10, leaving the PMN fraction. This fraction contains mostly neutrophils together with a small percentage of basophils and eosinophils. We refer to PMNs when the entire fraction is used in functional assays and to neutrophils when they are identified by flow cytometry.

B-cell enrichment

PBMCs were isolated from blood (EDTA) by initially spinning blood at 2500rpm (600g) for 15 min and collecting the buffy coat layer. This layer was then diluted with PBS, layered onto a Ficoll-Paque gradient, and centrifuged for 25 minutes at 2500rpm. Cells were washed with PBS, red blood cells were lysed with ACK lysing buffer (Lonza), washed again with PBS, and resuspended in R10. PBMCs or co-cultures of B-cells and PMNs were enriched for B-cells by depleting CD2⁺, CD4⁺, and CD14⁺cells using Miltenyi MACs following the manufacturer’s instructions.

Neutrophil Phagocytosis Assay:

Briefly, HIV gp120 was biotinylated with the Biotin-XX Microscale Protein Labeling Kit (Life Technologies, Grand Island, NY), and 3–5 μg of biotinylated gp120 was incubated with a 100-fold dilution in RPMI of 1μm Yellow-Green streptavidin-fluorescent beads (Life Technologies) for 25 min at RT in the dark. Autologous plasma (1:100 dilution) was added to >50,000 isolated PMNs in wells of a 96-well U-bottom plate. The bead-gp120 mixture was further diluted 5-fold in R10 and 50 μl was added to the cell/serum mixtures and incubated for 3 h at 37°C. Cells were fixed by adding 50 μL 2% PFA and assayed for fluorescent bead uptake by flow cytometry using a BD Biosciences LSRII. The phagocytic score of each sample was calculated as follows: (% phagocytosis x MFI)/10⁶. The values were standardized to background values (cells and bead only without serum) by dividing the phagocytic score of the test sample by the phagocytic score of the background sample.

Neutrophil Reactive Oxygen Species (ROS) quantification

Isolated PMNs (>2.5×10⁵) in 1mL RPMI were aliquoted into two wells of a 12 or 24 well plate. 50mM tert-butylhydroperoxide (tBHP), or luperox (Sigma) was diluted 1:500 into one of the wells containing neutrophils to stimulate reactive oxygen species production and incubated at 37˚C for 90min. CellROX Green (Invitrogen) was then added to each well for a final concentration of 5μM per well, and then incubated at 37˚C for 30min. Cells were then washed with PBS and fixed in 2% formaldehyde prior to quantifying the ROS using a BD LSRII flow cytometer.

Serum Luminex assay

Luminex data was acquired using serum with the Cytokine 29-Plex Monkey Panel (Invitrogen) according to the manufacturer’s specifications using a Bio-plex 200 system (Bio-Rad) that was properly validated and calibrated.

Neutrophil-B-cell co-culture

Autologous isolated PMNs were added to enriched B-cells at a 7:1 ratio (calculated from the number of enriched B-cells obtained) in 1mL in a 24-well plate. The cells were co-cultured for 18, 42, and 66 hrs in R10 prior to flow cytometry staining and analysis. Viability for B-cells following 66 hours of co-culture was roughly 80%, while PMNs were 40–50%. When used, IL-10 (Peprotech) was added at a concentration of 100ng/mL. The B-cell stimulation cocktail contained 0.1 μg/ml CpG (ODN-2006) (Operon), 0.5 μg/ml recombinant human sCD40L (Peprotech), and 50 ng/ml recombinant human IL-21 (Peprotech) (24). Supernatants were frozen, and later used for ELISA quantification of IgA, IgM, and IgG. 24-well transwell plates were used where indicated to separate PMNs (top) from B-cells (bottom).

Quantification of Igs by ELISA

Antibodies in culture supernatants were measured as follows: wells of Greiner high-binding ½ area 96-well plates were coated overnight at 4°C with 100 ng/well of SIV_M766or SIV_CG7VgD-gp120, SIV_mac251gp120, or 50ng/well of unconjugated Ig (Alpha Diagnostics) in 100 μl of sodium bicarbonate buffer (pH 9.6) (Sigma-Aldrich, St. Louis, MO). Wells were blocked with 200 μl of 1% BSA diluent/blocking solution (KPL) in distilled water for 2 h at RT. Culture supernatants (50 μl) were added and incubated for 1 h at 37°C. Plates were washed 5 times with 1X wash solution (KPL). Horseradish peroxidase-labeled goat anti-monkey Igs (50 μl at a 1:10,000 dilution; Alpha Diagnostics) was added, and plates were incubated for 1 h at 37°C. After washing as described above, 50 μl of 3,3′,5,5′-tetramethylbenzidine (TMB) peroxidase substrate (KPL) was added for 15–30 min at RT. Color development was stopped with 50 μl 1M Phosphoric acid (Sigma), and plates were read at 450 nm by using Biotek plate reader. SIV Env-specific antibody levels were expressed as ng Env-specific Ig/μg total Ig.

RNA isolation and cDNA Preparation

B-cells only or B-cells plus PMNs were cultured for 18 hours as described above. B-cells were collected from individual wells, positively selected for using PE-CD19 (Beckman Coulter) and anti-PE magnetic beads (Miltenyi) and lysed by adding RLT buffer. RNA was extracted using the AllPrep DNA/RNA micro kit (Cat: 80284, Qiagen, Valencia, CA, USA). Each homogenized cell lysate was transferred to an Allprep DNA spin column, and the flow-through was collected for RNA purification. An equal volume of 70% ethanol was added to the flow-through, mixed by pipetting, and transferred to an RNase minElute spin column. The column flow-through was discarded and the column was washed with RW1 buffer. The spin column was further washed by RPE buffer followed by 80% ethanol. Finally, RNase free water was added to elute the RNA. RNA was measured by nanodrop and 25 ng RNA was used for cDNA preparation with the RT2 First Strand kit (Cat No. 330404. Qiagen, Valencia, CA, USA). The RNA was treated with genomic DNA elimination mix for 5 minutes at 42^˚C followed by 1 minute incubation on ice. Equal amounts of reverse-transcription mix and the RNA/genomic DNA elimination mix were gently mixed by pipetting and incubated at 42^˚C for 15 minutes. The reaction was stopped by incubating at 95^˚C for 5 minutes. The first strand cDNA was diluted with RNase free water and used for Real-Time PCR.

Real-Time PCR

Relative transcription profiles of 84 cytokine and chemokine genes were determined by qRT-PCR using the RT²Profiler™ PCR Array Rhesus Macaque Cytokines and Chemokines Assay kit (Qiagen, Valencia, CA, USA). In addition, transcription of the housekeeping genes ACTB, B2M, GAPDH, LOC709186, and RPL13A was determined using specific primers in the kit. Real-time PCR reactions were set up in duplicate for each of the cytokines and the housekeeping genes. Amplification conditions were identical for all reactions and consisted of: 10 min at 95 °C, 40 cycles of 15 s at 95°C and 60 s at 60°C. Reaction samples had a final volume of 25 μl consisting of 12.5 μl of RT² SYBR Green qPCR Mastermix (Cat No./ID: 330500, Qiagen, Valencia, CA, USA) and 12.5 μl of cDNA. Amplifications were run in an ABI Prism 7500 Sequence Detection System (Applied Biosystems). Expression level differences were assessed using the ΔΔCt method.

Statistics:

Mann-Whitney-Wilcoxon test, repeated measures ANOVA, and Spearman correlations were used where indicated. For Fig. 3–5, Wilcoxon rank test, and Kruskal -Wallis tests were used as indicated in figure legends. The Jonckheere-Terpestra test was used on Fig 4M-O. In Fig. 7–8, p values were corrected for multiple comparisons by the Hochberg method. Statistics were generated using Graphpad Prism.

Study Approval:

All animal experiments were approved by Institutional Animal Care and Use Committees prior to study initiation. During the course of this study, rhesus macaques were housed in two facilities, each of which approved the work (Advanced BioScience Laboratories, Inc. (ABL), Rockville, MD, Protocol No. AUP526; and the NCI Animal Facility, Bethesda, MD, Protocol No. VB012). Each of these facilities is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. The standard practices closely follow recommendations made in the Guide for the Care and Use of Laboratory Animals of the United States—National Institutes of Health. Animals were housed in accordance with the recommendations of the AAALAC Standards and with the recommendations in the Guide for the Care and Use of Laboratory Animals. When immobilization was necessary, the animals were anesthetized with approximately 10 mg/kg of ketamine hydrochloride injected intramuscularly. All efforts were made to minimize discomfort of all animals used in the study, including provision of peri-operative and post-operative analgesia and strict accordance to humane endpoint criteria. Details of animal welfare and steps taken to ameliorate suffering were in accordance with the Guide and the recommendations of the Weatherall report, “The use of non-human primates in research”, as approved by the relevant IACUCs.

Results:

Neutrophil tissue comparison

Neutrophils are the most abundant white cell in the blood, but also are able to rapidly migrate to tissues in response to injury or infection. As an initial site of HIV/SIV infection is the rectal mucosa, with subsequent systemic spread of the virus to lymph nodes (LNs), we initially phenotypically compared neutrophils from these sites. Neutrophils obtained from whole blood, rectal pinches (RP), and LNs of naïve rhesus macaques were phenotypically analyzed by flow cytometry from these respective tissues (Fig. 1) and defined as CD66⁺(Supplemental Fig. 2). The presence of eosinophils cannot be discounted in light of the antibody panel used. Eosinophils most likely represent 5% or less of the CD66⁺cells, based on ongoing experimentation not presented here. CD15 and CD16, markers typically used for identification of human neutrophils, could not be used, as there is no working CD15 antibody clone for rhesus macaques, and CD16 is not expressed by rhesus macaque neutrophils (25). We observed more neutrophils in the blood compared to both the RP and the LN, as expected (Fig. 1A). In Fig. 1B through L, the frequencies of CD66⁺neutrophils expressing various markers are shown. CCR7 expression levels were significantly elevated in LN compared to blood (Fig. 1B), confirming the observation that CCR7 is involved in migration of neutrophils to LNs in mice (26). In contrast, CXCR2 (IL-8Rβ) expression, a neutrophil chemoattractant chemokine, was lower in blood than in RP, presumably a more inflammatory site (Fig. 1C). Several other markers were expressed more abundantly in tissue neutrophils compared to blood including TNFα, associated with activation, which exhibited significantly higher frequency in RP neutrophils than blood, with an intermediate level in LN (Fig. 1D). Additionally, there were almost no PD-L1 expressing neutrophils associated with suppression in blood, but significantly more in RP (Fig. 1E). Neutrophils associated with provision of B-cell help were also more abundant at tissue sites. BAFF (BLyS) expressing neutrophils were significantly lower in the blood relative to RP and LN (Fig. 1F), and IL-21-positive neutrophils were significantly lower in blood than RP, and lower than LN (Fig. 1G). In contrast, a number of markers associated with degranulation, phagocytosis, and NET formation, were elevated in blood neutrophils compared to those at tissue sites. Neutrophils expressing CD11b, present in neutrophil granules, were more prevalent in blood than in RP or LN (Fig. 1H). However, expression of myeloperoxidase (MPO), also present in neutrophil granules, showed an increasing abundance in LN and RP respectively (Fig. 1I). Fcγ receptors CD32 and CD64, important for phagocytosis, were more highly expressed by neutrophils in the blood than in both RP and LN (Fig. 1J-K), with blood neutrophils exhibiting significantly greater expression than RP for both. The abundance of neutrophils expressing PSGL-1 (CD162) associated with NET formation (27), was greater in blood than LN, and in LN than RP (Fig. 1L), with blood significantly greater than RP. Overall, circulatory neutrophils exhibited a distinctly different phenotype when compared to those from LN and RP, with those found in blood significantly distinct from neutrophils found in RPs. LN-resident neutrophils exhibited mostly intermediate properties compared to RP and blood. While we continued by examining neutrophils in blood only, it is clear that the cells at the three sites all exhibited a phenotype consistent with functional neutrophils.

Figure 1. Neutrophil phenotypes in whole blood, LN, and RP of rhesus macaques.

Neutrophils from naïve animals (n=6) from whole blood, LN, and RP were stained and analyzed by flow cytometry. All panels shown are gated on neutrophils (Live cells, singlets, CD2-, CD20-, CD66+). (A) Percent total neutrophils; (B-L) sequential plots of percent positive neutrophils for CCR7, CXCR2, TNFα, PD-L1, BAFF, IL-21, CD11b, MPO, CD32, CD64, and PSGL-1. Statistical differences were determined using repeated measures ANOVA, p values represented by *, <0.05; **, <0.01.

Neutrophils are stimulated by replicating Adenovirus-recombinant vaccination and exhibit a phenotype consistent with facilitating B-cell help

In the context of a mucosal prime-systemic boost vaccine strategy, we evaluated the response of neutrophils in rhesus macaques over the course of vaccination and post-SIV infection. The vaccination protocol is outlined in Supplemental Figure 1along with the timepoints of sample collection. Two weeks after the second Ad5hr-recombinant administration, whole blood staining was used to analyze the neutrophil response to the vaccination. A significant decrease (p<0.0001) in the percentage of CD66⁺neutrophils was observed (Fig. 2A). Significant increases occurred in both the frequency of TNFα⁺neutrophils (p=0.040) and the mean fluorescence intensity (MFI) of TNFα expression (p=0.0054) (Fig. 2B-C). BAFF (BLyS) and IL-21 were assayed on neutrophils to address their propensity to mediate B-cell help, and their respective MFI’s were significantly changed (Fig. 2E, G). BAFF MFI decreased (p=0.0005), while IL-21 MFI increased (p=0.0002), and the percentages of neutrophils expressing each molecule exhibited non-significant increases (Fig. 2D, F). CD11b and myeloperoxidase (MPO) were used as markers for the ability of neutrophils to degranulate. The respective percentages of CD11b⁺and MPO⁺neutrophils both decreased significantly (p=0.0014 and p=0.018 respectively) (Fig. 2H, J), while the MFI’s of both expressed molecules significantly increased (p=0.0037 and p<0.0001 respectively) (Fig. 2I, K). Gating for these markers is shown in Supplemental Fig. 3. These results indicated that the Ad5hr priming immunizations had a significant effect on neutrophils in the blood, causing stimulation or maturation, greater levels of BAFF and IL-21 expression intimating a propensity for B-cell help, and enhanced expression of degranulation markers, overall suggesting greater functional capacity.

Figure 2. Neutrophil responses to replicating adenoviral vaccination.

Neutrophils from whole blood were analyzed by flow cytometry at a pre-vaccination timepoint (n=33) and two-weeks after the 2^ndadenovirus vaccination (n=34). All panels shown are CD66⁺neutrophils (Live cells, singlets, CD2-, CD20-). (A) Percent total neutrophils; (B-K) sequential plots of percent positive neutrophils and MFI for TNFα, BAFF, IL-21, CD11b, and MPO. Statistical differences were determined using the Mann-Whitney-Wilcoxon test (B-C, E) and repeated measures ANOVA (A, G, I-K).

Neutrophil response to systemic vaccination is primarily a response to adjuvant

As neutrophils respond rapidly to stimuli in providing a first line of innate immune defense against pathogens (28), we examined blood neutrophils at 3 days as well as 2 weeks following the second systemic boost with either ALVAC/Env or DNA&Env. The neutrophils were analyzed by whole blood fluorescent antibody staining. For all four functional markers examined: TNFα, IL-21, CD11b, and MPO, the patterns of response were identical across the 3 immunization groups (ALVAC/Env, DNA&Env and Controls) as shown by analysis of the frequencies of neutrophils expressing the various markers between the timepoints 2 weeks after the second priming immunization and 3 days after the second booster immunization (Fig. 3A– D). As the control macaques received only alum, it appears the neutrophil response was primarily to the adjuvant rather than the SIV immunogens administered. Highly significant increases in the frequencies of neutrophils expressing the four different markers were seen in the ALVAC/Env group, in part attributable to the lower frequencies seen at the 2 week post-second Ad time point. The transient nature of the response to the booster immunogens was evident, as was also seen when comparing the post-boost responses with pre-immunization values (Supplemental Fig. 4). By 2 weeks following the second boost, only the frequencies of TNFα⁺neutrophils in the DNA&Env group of macaques were still significantly elevated above the pre-immunization values.

Figure 3. Neutrophil responses to systemic boosts with ALVAC/Env and DNA&Env.

Neutrophils from whole blood were analyzed by flow cytometry both three days and two weeks after a second boost with either ALVAC/Env (n=12), DNA&Env (n=14), or alum only (control group, n=8) and compared to phenotypes of neutrophils obtained 2 weeks after the second immunization with the Ad5hr vector. All graphs display CD66⁺neutrophils (Live cells, singlets, CD2-, CD20-). (A) The frequencies of neutrophils expressing TNFα increased significantly between the 2^ndAd+2 wks and the 2^ndboost + 3d timepoints (# in 3 groups combined: p = 0.0025 by the Wilcoxon signed rank test). (B-D) Similarly, the frequencies of neutrophils expressing IL-21, CD11b, and MPO increased between the same time points (¥ in 3 groups combined: p = 0.0001 in each panel). (A-D) Changes in frequencies of neutrophils expressing TNFα, IL-21, CD11b, and MPO between the indicated timepoints in the ALVAC group compared to the other two macaque groups were determined by Kruskal-Wallis tests (p values represented by *, <0.05; ***, <0.001; and ****, <0.0001). (E) The percentage of CD66⁺neutrophils 2 weeks post 2^ndboost negatively correlated with the number of challenges needed for infection. (F)The percentage of neutrophils expressing TNFα 3d post 2^ndboost negatively correlated with the viral load 2 weeks post-infection. (E-F) Statistics generated using Spearman correlation.

Further analysis showed that at the 2 weeks post 2^ndboost timepoint, there was a negative correlation between mild neutropenia and the number of challenges needed for infection (Fig. 3E), possibly indicating that a neutrophil reaction to vaccination that results in neutrophil death is beneficial. Alternatively, the neutropenia might simply reflect migration of neutrophils from the blood to the rectal site of SIV exposure where they effectively combat the potential infection. There was also a negative correlation between the frequency of TNFα+ neutrophils 3 days following the second booster immunization and the viral load 2 weeks post-infection (Fig. 3F), indicating a sustained positive influence of TNFα. Collectively, these data highlight the innate role of neutrophils in response to vaccination and their potential to influence the outcome of viral exposure.

Neutrophil response during acute and chronic infection

It has been previously reported that neutrophil dysregulation occurs during HIV infection (28–31). We sought to determine what dysfunction occurs following SIV infection relative to neutrophil activation, ability to contribute to B-cell help, and functionality. The frequency of CD64⁺(FcγRI) neutrophils significantly decreased in blood both during acute (two weeks post-infection) and chronic infection (16–40 weeks post-infection) when compared to neutrophils post boost (Fig. 4A), and the MFI of CD64 during chronic infection was significantly lowered in comparison to both post boost and acute infection timepoints (Fig. 4B). The percentage of TNFα⁺neutrophils significantly decreased during both acute and chronic infection (Fig. 4C), but TNFα MFI, although exhibiting a significant rebound from acute to chronic infection timepoints, was unchanged over the course of infection when compared to the post-boost time point (Fig. 4D). Although the frequency of BAFF⁺neutrophils did not change at the timepoints examined (Fig. 4E), BAFF MFI on neutrophils exhibited a significant increase at both timepoints post-infection compared to post-boost (Fig. 4F). The frequency of IL-21⁺neutrophils exhibited a significant decrease during both acute and chronic infection (Fig. 4G), whereas IL-21 MFI expression was unchanged (Fig. 4H). The percentage of CD11b⁺neutrophils exhibited a significant decrease during the acute infection stage (Fig. 4I), with CD11b MFI significantly increasing during chronic infection compared to the pre-infection level (Fig. 4J). MPO⁺neutrophil frequency and MPO MFI were both significantly decreased during acute infection (Fig. 4K-L). The percentage of MPO⁺neutrophils during chronic infection remained below pre-infection levels (Fig. 4K), whereas there was a significant rebound in MPO MFI levels during the chronic infection period (Fig. 4L). The observed changes in the neutrophil population reflect some dysregulation, with decreased maturation and stimulation markers, B-cell help markers (IL-21), and granule markers.

Figure 4. Neutrophil responses following infection by SIV_mac251.

Neutrophils from whole blood were analyzed by flow cytometry 2 weeks after the 2^ndboost (n=26), two weeks after infection (acute, n=17) with SIV_mac251, and at a chronic timepoint (16–40 weeks post infection, n=17). All graphs display CD66⁺neutrophils (Live cells, singlets, CD2-, CD20-). (A-L) sequential plots of percent positive neutrophils and MFI for CD64, TNFα, BAFF, IL-21, CD11b, and MPO. (M-O) Compared to the number of challenges needed for infection, from left to right there is a negative correlation with the total number of CD66+ neutrophils, a positive correlation with IL-21 MFI, and a positive correlation with TNFα MFI. Statistics generated using Wilcoxon signed rank test (A-E, I), repeated measures ANOVA (F-H, J-L), and the Jonckheere-Terpstra test (M-O).

Mild neutropenia in the blood, seen during acute infection, was again significantly associated with an increased number of SIV challenges needed for infection (Fig. 4M). Positive correlations were seen between the number of challenges and both IL-21 and TNFα MFI at the 2 weeks post-infection timepoint (Fig. 4N-O). It is possible that these latter associations reflect neutrophil responses to multiple SIV exposures, which continued into the acute infection period.

Frequencies of neutrophils expressing various phenotypic markers correlate with serum cytokine and chemokine levels over the course of immunization

Serum cytokines and chemokines were also measured at the same timepoints that neutrophils were phenotypically analyzed, and their expression was examined for association with various neutrophil subsets as summarized in Table I. Two weeks following the second Ad5hr-recombinant vaccination, IL-10 and HGF expression correlated with the frequency of MPO⁺neutrophils. Three days after the 2^ndboost, expression of IL-6 and MIF correlated with the frequency of IL-21⁺and MPO⁺subsets, respectively. Two-weeks following the 2^ndboost, IL-1β expression correlated significantly with the overall percentage of neutrophils (CD66⁺) present, while at the same time point, G-CSF, IL-10, MIP-1α, GM-CSF, MIP-1β, EGF, IFNγ, MIF, IL-1Rα, and IL-2 serum levels correlated with the frequency of MPO⁺and/or CD64⁺neutrophil phenotypes. The percentages of neutrophils expressing CD64 and MPO was clearly related to some degree with serum cytokines, as correlations between the two neutrophil phenotypes were seen with 13 separate chemokines/cytokines. The abundance of correlations between serum cytokine levels and neutrophil subsets clearly suggests that neutrophils interact significantly with the systemic immune system.

Table I.

Correlations of neutrophil phenotypic subsets with expression of serum cytokines.

Cytokine	Time point	CD66+	IL-21+	MPO+	CD64+
IL-10	2 wk post-2ndAd			r = −0.58 p = 0.023
	2 wk post-2ndboost				r = −0.50 p = 0.025
HGF	2 wk post-2ndAd			r = −0.50 p = 0.042
IL-6	3 d post-2ndboost		r = 0.57 p = 0.035
MIF	2 wk post-2ndboost			r = −0.49 p = 0.043
	2 wk post-2ndboost			r = −0.51 p = 0.042
IL-1β	2 wk post-2ndboost	r = −0.56 p = 0.047
G-CSF	2 wk post-2ndboost				r = −0.55 p = 0.019
MIP-1α	2 wk post-2ndboost			r = −0.53 p = 0.039	r = −0.54 p = 0.020
GM-CSF	2 wk post-2ndboost				r = −0.48 p = 0.032
MIP-1β	2 wk post-2ndboost				r = −0.50 p = 0.025
EGF	2 wk post-2ndboost				r = −0.47 p = 0.035
IFNγ	2 wk post-2ndboost				r = −0.51 p = 0.025
IL-1Rα	2 wk post-2ndboost			r = −0.52 p = 0.039
IL-2	2 wk post-2ndboost				r = −0.58 p = 0.019

Neutrophils were phenotyped as CD66⁺, and subsets were further phenotyped as IL-21⁺, MPO⁺and CD64⁺. Serum cytokine levels were obtained by Luminex assay as described in Methods. Significant correlations between neutrophil frequencies and cytokine levels were obtained by Spearman correlation analysis. Twenty-seven cytokines were tested; each column was corrected by the FDR method, p≤0.05.

PMN oxidative burst capacity

PMNs were isolated 2 weeks following both the second Ad5hr-recombinant prime and second boost, as well as during chronic infection (>16 weeks post infection) and assayed for their functional capacity. They were either untreated or stimulated with tBPH (see methods), prior to being stained for quantification of ROS. Following the 2^ndAd5hr-recombinant immunization, there was a significant increase in the capacity to generate ROS in response to stimulation, when compared to the pre-vaccination timepoint (Fig. 5A). Following the second boost, there was a significant decrease in the capacity to produce ROS in the ALVAC/Env immunized animals (Fig. 5B), but no change in the DNA&Env group in comparison to the post-priming levels. Interestingly, ALVAC animals could better generate ROS prior to the 2^ndboost, and DNA vaccinated animals could better generate ROS following the 2^ndboost. In this study, seven vaccinated macaques resisted 15 weekly SIV intrarectal exposures (Musich et al., in preparation), and we have considered them protected. In comparing ROS production between the chronically infected animals and the protected animals, there was a clear impairment of the functionality of the PMNs from infected animals (Fig. 5C). Overall, there was a clear improvement following Ad5hr priming, and a clear impairment during chronic infection, of the capacity of PMNs to produce ROS in response to stimulation.

Figure 5. Neutrophil ability to generate reactive oxygen species in response to stimulation following adenoviral prime, ALVAC/Env and DNA&Env boost, and SIV_mac251infection.

PMNs were isolated from whole blood and incubated with and without Luperox for 1hr, prior to incubation with CellROX for 30min. Cells were fixed and ROS was then quantified by flow cytometry. (A) Pre vs. 2 weeks post 2^ndadenovirus vaccination (n=26). (B) 2 weeks post 2^ndadenovirus vaccination vs. 2 weeks post 2^ndboost, broken down by immunization groups. (ALVAC n=12, DNA n=14, Controls n=8) (C) Animals that exhibited protection from challenge (n=7) vs. chronically infected animals (n=17). Statistics generated using the Wilcoxon signed rank test (A and B [within ALVAC]) and the Kruskal-Wallis test (B [ALVAC vs. DNA] and C).

PMN Phagocytic Capacity

To further evaluate the effect of vaccination on the functional capacity of neutrophils, their ability to phagocytose specific antigens was quantified. PMNs were isolated from blood at the 2 weeks post 2^ndboost timepoint and incubated with antigen-coated fluorescent beads and autologous plasma prior to quantification of bead uptake using flow cytometry. Geometric mean binding antibody titers at this time point against SIV_CG7VEnv were 1.43 X 10⁶for the ALVAC/Env group and 1.21 X 10⁶for the DNA&Env group. Titers against SIV_M766Env were 9.64 X10⁵and 1.96X10⁶for the ALVAC/Env and DNA&Env groups, respectively. Titers of the control macaques were all <50. For all three gp120 antigens used from SIV strains CG7V, M766, and Mac251, PMNs from vaccinated macaques had significantly greater phagocytic capacity compared to controls (Fig. 6A). SIV_CG7VEnv-specific serum antibody titers exhibited a positive correlation with phagocytosis of both SIV_CG7Vand SIV_Mac251gp120-coated beads (Fig. 6B-C). These correlations indicate the phagocytic capacity observed in PMNs is derived from the SIV Env-specific antibody present systemically.

Figure 6. Neutrophil phagocytic capacity following systemic boosting.

PMNs were isolated from whole blood and incubated with fluorescent beads conjugated to the indicated SIV envelope protein (M766 or CG7V gD-gp120) or SIV_mac251gp120). Cells were fixed after 3hr incubation and phagocytosis was quantified by flow cytometry. (A) PMNs from vaccinated animals (n=26) exhibited significantly greater phagocytosis of SIV Env-coated beads compared to controls (n=8). (B) Phagocytosis of SIV_CG7VEnv-coated beads correlated with the SIV_CG7VEnv-specific antibody titer detected in serum at the same timepoint (n=34). (C) Phagocytosis of SIV_mac251Env-coated beads correlated with the SIV_CG7VEnv-specific antibody titer detected in serum at the same timepoint (n=34). (A) Statistics generated using the Mann-Whitney-Wilcoxon test. (B-C) Statistics generated using the Spearman rank test.

PMNs co-cultured with autologous B-cells elicit responses indicative of B-cell help

Blood taken from chronically infected, protected and naïve rhesus macaques was processed for enrichment of B-cells and autologous PMNs. The B-cells were then co-cultured either alone, stimulated with a cocktail to induce B-cell proliferation and differentiation into antibody-secreting cells, or with autologous PMNs, and longitudinally sampled after 18, 42, and 66 hours for analysis by flow cytometry. There was a clear induction of AID⁺B-cells when co-cultured with autologous PMNs, significantly more than B-cells alone or stimulated B-cells, in protected, infected, and naïve animals, when comparing the area under the curve (AUC) values (Fig. 7A-C). CD27 MFI also significantly decreased on B-cells co-cultured with PMNs compared to B-cells alone or stimulated B-cells (Fig. 7D-F). B-cells co-cultured with PMNs exhibited the lowest frequencies of IgD⁺cells compared to stimulated or unstimulated B-cells in both protected and infected animals (Fig. 7G-H) but not the naïve macaques (Fig. 7I). As has been previously described (32), co-culture with PMNs resulted in a loss of IgM⁺B-cells in all three groups of animals (Fig. 7J-L). Lastly, co-culture with PMNs led to the greatest frequencies of Ki-67⁺B-cells compared to unstimulated or stimulated B-cells in all three groups of animals (Fig. 7M-O). Taken together, PMN co-culture with autologous B-cells clearly induced class switching, seen in AID induction and loss of CD27 and IgD, and stimulation, seen by the increase in Ki67 expression.

Figure 7. PMNs co-cultured with autologous B-cells lead to responses indicative of B-cell help.

PMNs were isolated from whole blood of protected (n=4), chronically infected (n=4), and naïve macaques (n=4) and co-cultured with autologous PBMCs enriched for B-cells. Enriched B-cells were also cultured alone or with a stimulation cocktail in parallel. B-cells were analyzed by flow cytometry prior to co-culture, and 18, 42, and 66 hrs post-co-culture. % AID⁺B-cells from (A) protected, (B) infected, and (C) naïve animals. CD27 MFI on B-cells from (D) protected, (E) infected, and (F) naïve animals. % IgD⁺B-cells from (G) protected, (H) infected, and (I) naïve animals. % IgM⁺B-cells from (J) protected, (K) infected, and (L) naïve animals. % Ki67⁺B-cells from (M) protected, (N) infected, and (O) naïve animals. Area under the curve values (AUCs) calculated from either the raw data or the log- or arcsine-transformed values were consistent with the assumptions of repeated measures ANOVA. * and # indicates a significant difference between B-cells + PMNs and stimulated B-cells or B-cells alone respectively. All p values were p<0.0001 except (D) # (p=0.016), (G) # (p=0.0003), (M-O) * (p=0.019), and (M,O) # (p<0.0005). The p values were corrected for multiple comparisons by the Hochberg method.

In addition to assessing B cell help within each of the three subgroups of macaques, we also asked if there were differences in the responses of B cells to autologous PMNs between the naïve, chronically infected, and protected macaques by comparing AUC values for the B cell + PMN co-cultures across these groups. For the AUC analysis to be unbiased, the baseline values had to be similar. The % AID⁺and % IgM⁺responses could not be analyzed, as the naïve macaques had higher baseline AID⁺values and lower IgM⁺values than the other two groups by the Kruskal-Wallis test (p = 0.0032 and 0.019, respectively; p = 0.016 and p = 0.077 with the Hochberg correction for multiple comparisons). The different baseline values between the naïve and infected macaques most likely result from immune dysfunction in the SIV-infected animals. Whether the continued SIV exposures (15 challenges) of the protected macaques also resulted in some B cell dysfunction will require further study. AUC analysis of the CD27 MFI and % IgD⁺B cells suggested greater class switching in the protected and infected macaques. The naïve macaques had marginally higher AUC values in the CD27 MFI analysis compared to the protected animals (p = 0.044) and also had the highest AUC values for % IgD+ cells compared to the protected and infected macaques (p <0.0001).

Mechanism of PMN/B-cell interaction involves cell-cell contact with little influence of IL-10

After determining that PMNs elicited a response in B-cells, we investigated the mechanism of this interaction. It has been previously reported that IL-10 can enhance the response of B-cells to neutrophils (17), and we attempted to confirm this observation in the rhesus macaque model. Additionally, by separating PMNs from B-cells using a transwell system, we investigated whether it was largely soluble factors contributing to the B-cell help or if it was contact driven (Fig 8A-L). The percentage of B-cells expressing CD69, an early activation marker, increased over time when co-cultured with PMNs, both with and without IL-10, in protected and naïve animals (Fig. 8A, C). In the naïve group interestingly, B-cells cultured with PMNs and IL-10 had significantly greater CD69 expression than all other B-cells (Fig. 8C). No difference in the frequency of these cells was observed in any group between B-cells cultured alone and B-cells co-cultured with PMNs separated by a transwell (Fig. 8A-C). CD80⁺B-cells had a similar response to CD69⁺B-cells when co-cultured with PMNs in protected and naïve groups (Fig. 8D, F), but no significance was observed. There was a loss of IgD⁺B-cells in protected and naïve animals when cultured with PMNs with or without IL-10 (Fig. 8G, I), compared to B-cells cultured alone or in a transwell with PMNs. Similarly, in protected and naïve animals, the IgG⁺B-cell frequency increased significantly when co-cultured with PMNs with or without IL-10 compared to B-cells cultured alone or with PMNs separated by a transwell (Fig. 8J, L). These results again illustrate that PMNs stimulate B-cells, and cause class switching and Ig production. The addition of IL-10 to the co-culture of B-cells and PMNs gave varied results and requires further evaluation. The effect of PMNs on B-cells in this experiment was contact-dependent as shown by the lack of response when the cells were separated by a transwell. In every panel of Fig. 8, the responses of B cells + PMN compared to B cells alone were not different when the cells were separated in transwells.

Figure 8. PMNs co-cultured with autologous B-cells require cell-cell contact to cause responses indicative of B-cell help.

PMNs were isolated from whole blood of protected (n=4), chronically infected (n=4), and naïve macaques (n=4) and co-cultured with autologous PBMCs enriched for B-cells with and without IL-10, or with enriched B-cells separated by a transwell system. Enriched B-cells were also cultured alone in parallel. B-cells were analyzed by flow cytometry prior to co-culture, and 18, 42, and 66 hrs post-co-culture. % CD69⁺B-cells from (A) protected, (B) infected, and (C) naïve animals. % CD80⁺B-from (D) protected, (E) infected, and (F) naïve animals. % IgD⁺B-cells from (G) protected, (H) infected, and (I) naïve animals. % IgG⁺B-cells from (J) protected, (K) infected, and (L) naïve animals. Area under the curve values (AUCs) calculated from either the raw data or the log- or arcsine-transformed values were consistent with the assumptions of repeated measures ANOVA. In (C) # shows B+PMN+IL-10 is significantly greater than both B-cell only and B+transPMN (p<0.0001), and $ shows B+PMN+IL-10 is significantly greater than B+PMN (p=0.0018). In (D), * indicates B+PMN is significantly greater than B-cell only (p=0.040). & shows a significant difference between B+PMN +/− IL-10 and B-cells only and B+transPMN in (G, I) p<0.015 and in (J, L) p<0.035. The p values were corrected for multiple comparisons by the Hochberg method.

We again compared B cell responses across the 3 subgroups of macaques by AUC analysis of the B cell + PMN co-cultures. Comparisons of the % CD69⁺and % IgG⁺B cells could not be made as the baseline values differed. The % CD69⁺baseline values ranged from the lowest in the infected group to the highest in the protected group (p = 0.0036, p = 0.014 after Hochberg correction), perhaps reflecting SIV-induced dysfunction in the infected macaques. The naïve group had the lowest % IgG⁺baseline values (p = 0.021; p = 0.064 after correction), not surprising as the protected and infected macaques had long-term exposures to both vaccine and challenge antigens resulting in a greater IgG⁺cell frequency. The AUC values for % CD80⁺cells were not different across the three macaque groups. The naïve macaques had the lowest AUC values for % IgD⁺cells compared to both the protected and infected macaques (p = 0.0006 and p = 0.0009, respectively), suggesting the greatest functionality. Notably, the three randomly selected naïve macaques studied here were different from those studied in the experiment of Figure 7, and their B cells were much more responsive to the help provided by the autologous PMNs.

B-cells co-cultured with PMNs produce IgG, IgA, and IgM

In addition to determining if antibodies were produced by B-cells cultured with PMNs using flow cytometry to detect their presence in the cell membrane, the co-culture supernatants were also assayed by ELISA to determine if the PMN-stimulated B-cells secreted antibodies. We initially ascertained by flow cytometry that B-cells co-cultured with PMNs in protected and chronically infected but not naïve animals exhibited elevated levels of plasmablasts, CD27^-CD20^-HLA-DR⁺CD38⁺B-cells (Fig. 9A-C), which have been shown to secrete antibodies in macaques in response to vaccination (33). The co-culture supernatants from these chronically infected and protected animals had increased levels of SIV Env-specific IgG when comparing their AUCs to B-cells alone, although statistical significance was not ascertained as there were too few samples for non-parametric tests, and the ANOVA model was not appropriate due to the B-cell only group having all zero values (Fig. 9E, G). Elevated levels of IgM antibodies were not observed (Fig. 9D, F). Specific IgA was only detected in the co-culture supernatant from infected animals (Fig. 9H). As seen in the flow cytometry results, the PMNs provided help to B-cells as seen by the increased antibody secretion compared to B-cells alone, and the addition of IL-10 did not have an effect in this experiment. Surprisingly, however, co-culture of B-cells with PMNs in a transwell system did not ablate antibody secretion. Levels of secreted antibody were similar to those in supernatants of co-cultures where the cells were not separated. Thus, rather than contact dependence, help by neutrophils for B-cell antibody secretion was attributable to a soluble factor. The difference seen in the neutrophil help provided is discussed below.

Figure 9. PMNs co-cultured with autologous B-cells stimulate generation of plasmablasts and antibody secreting cells.

PMNs were isolated from whole blood of protected (n=4), chronically infected (n=4), and naïve macaques (n=4) and co-cultured with autologous PBMCs enriched for B-cells with and without IL-10, as well as separated by a transwell system. B-cells were harvested along with supernatants prior to co-culture and at 18, 42, and 66 hrs post-co-culture. Supernatants were analyzed for SIV-specific Igs by ELISA, and B-cells were analyzed by flow cytometry. The percentage of HLA-DR⁺CD38⁺B-cells of the CD27- CD20- cells are shown for (A) protected, (B) chronically infected, and (C) naïve animals. SIV_mac251-specific antibody relative to total antibody isotype in supernatants collected from protected animal co-cultures is shown for (D) IgM and (E) IgG, and in supernatants from chronically infected animal co-cultures for (F) IgM, (G) IgG, and (H) IgA. * indicates p<0.0015 for B+PMN AUC vs. B-cell only and Btrans+PMN. # indicates p<0.0025 for B+PMN+IL-10 AUC vs. B-cell only and Btrans+PMN. Statistics were generated using repeated measures ANOVA.

A comparison of B cell responses by AUC analysis of B cell + PMN co-cultures of the % HLA-DR⁺CD38⁺cells across the three macaque subgroups was not appropriate as the baseline values were significantly different, with the chronically infected group having the highest values (p = 0.025). This is not surprising as the chronically infected macaques were continually exposed to SIV antigens, leading to a greater frequency of plasmablast cells. A comparison of IgM and IgG secretion of SIV-specific antibodies in the B cell + PMN co-cultures revealed no difference in AUC values for the two IgM subgroups (Fig. 9D, F) while the chronically infected macaques exhibited a higher AUC value for IgG secretion compared to the protected macaques (Fig. 9E, G; p = 0.0040), again attributable to greater antigen exposure. Taken together, the comparisons of B cell + PMN co-cultures across the three macaque subgroups (Figs. 7–9) are consistent with some dysfunction in macaques exposed to SIV, while at the same time those macaques continually exposed to antigen exhibit a greater potential for antibody production.

B-cells co-cultured with PMNs have differential gene expression compared to B-cells alone

B-cells were positively selected from co-culture with PMNs as described in Methods and compared to autologous B-cells cultured alone by RT-PCR array. Out of 84 genes assayed, 28 were upregulated more than 2.5-fold when B-cells were co-cultured in the presence of PMNs (Table II). This includes genes associated with proinflammatory cytokines, antibody production, differentiation, and chemotaxis/migration. The gene expressing IL-8 was most highly upregulated, followed by CCL21, CXCL1, and RANKL. Expression of CD40 ligand, IL-21, and BAFF genes which contribute to antibody production were all upregulated. There was a clear influence of PMNs on B-cells, including alteration of gene expression.

Table II.

Cytokine and chemokine gene expression of B cells co-cultured with PMNs compared to B cells alone.

Gene category	Gene Symbol	Gene title	Fold Up-regulation
Proinflammatory cytokines	IL-1α	Interleukin 1, alpha, hematopoietin 1	3.33
	IL-1β	Interleukin 1, beta, leukocytic pyrogen, leukocytic endogenous mediator, mononuclear cell factor, lymphocyte activating factor	3.15
	IL-23A	Interleukin 23, alpha subunit p19	3.12
	IL12A	Interleukin 12A (IL-12 p35)	3.08
	IL17F	Interleukin-17F-like (ML-1)	3.06
	IL12B	Interleukin 12B (natural killer cell stimulatory factor 2, cytotoxic lymphocyte maturation factor 2, p40)	2.98
	IL6	Interleukin 6 (interferon, beta 2)	2.71
Anti-inflammatory	OSM	Oncostatin M	2.94
	IL-10	Interleukin 10, cytokine synthesis inhibitory factor (CSIF)	2.92
Pro and Anti Inflammatory	IL27	Interleukin-27 subunit alpha-like	3.05
Antibody Production	CD40LG	CD40 ligand	3.15
	IL-21	Interleukin 21	3.06
	TNFSF13b	Tumor necrosis factor (ligand) superfamily, member 13b, B-cell activating factor (BAFF)	2.76
Differentiation	TNFSF11	Tumor necrosis factor (ligand) superfamily, member 11, Receptor activator of nuclear factor kappa-B ligand (RANKL)	3.52
	BMP4	Bone morphogenetic protein 4	3.36
	NODAL	Nodal homolog (mouse)	3.03
	CSF3	Colony stimulating factor 3 (granulocyte), G-CSF or GCSF	2.84
Chemotaxis and Migration	CXCL8	Interleukin 8	7.84
	CCL21	Chemokine (C-C motif) ligand 21	3.75
	CXCL1	Chemokine (C-X-C motif) ligand 1, melanoma growth stimulating activity alpha, neutrophil-activating protein 3 (NAP-3)	3.53
	CX3CL1	Chemokine (C-X3-C motif) ligand 1, fractalkine	3.37
	CCL2	Chemokine (C-C motif) ligand 2, monocyte chemoattractant protein 1 (MCP1)	3.20
	CCL8	Chemokine (C-C motif) ligand 8, monocyte chemoattractant protein 2 (MCP2)	3.15
	CXCL5	Chemokine (C-X-C motif) ligand 5	3.06
	CCL18	Chemokine (C-C motif) ligand 18, Pulmonary and activation-regulated chemokine (PARC), dendritic cell (DC)-chemokine 1 (DC-CK1), alternative macrophage activation-associated CC chemokine-1 (AMAC-1), and macrophage inflammatory protein-4 (MIP-4)	2.76
	CXCL12	Chemokine (C-X-C motif) ligand 12, stromal cell-derived factor 1 (SDF1)	2.70
	MSTN	Myostatin, growth differentiation factor 8 (GDF8)	2.65
	CCL17	Chemokine (C-C motif) ligand 17, thymus and activation regulated chemokine (TARC)	2.62
	CXCL13	Chemokine (C-X-C motif) ligand 13, B lymphocyte chemoattractant (BLC) or B cell-attracting chemokine 1 (BCA-1)	2.45

Gene expression was determined by RT-PCR array as described in Methods. Genes are arranged in descending order of up-regulation within each category. Naïve animals were used, n = 3.

Discussion:

In this study we longitudinally examined the response of rhesus neutrophils to vaccination against, and infection with SIV. We utilized neutrophils from blood, showing them to be phenotypically distinct from those found in LN or mucosal tissue (Fig. 1). Those isolated from rectal pinches exhibited the highest expression of CXCR2 (IL-8Rβ) (Fig. 1C), suggesting that this is an important homing factor for the recruitment of neutrophils to mucosal sites, in agreement with observations made in mice (34). These mucosal neutrophils are able to contribute to immunosuppression via elevated PD-L1 expression (Fig. 1E) in response to high antigenic exposure in the intestine, as has been observed in other studies where significant antigenic stimulation was occurring (35–37). In the lymph node, which had the lowest density of neutrophils (Fig. 1A), CCR7 appeared to be important for neutrophil recruitment (Fig. 1B), as has been demonstrated previously (26, 38). Elevated levels of BAFF and IL-21 in the LN and the mucosa compared to the blood (Fig. 1F, G) most likely contribute to B-cell help in the tissues, although this was not tested in this study. Lymph node neutrophils overall had an intermediate phenotype compared to circulating and mucosal neutrophils. Circulating neutrophils were used here in all subsequent studies. In naïve macaques, representative of neutrophils in macaque samples obtained prior to immunization, they expressed high levels of Fc receptors (Fig. 1J-K), CD11b (Fig. 1H), and P-selectin glycoprotein ligand-1 (PSGL-1) (Fig. 1L), which are respectively involved in phagocytosis (39), degranulation (40), and NET formation (27). This functional propensity of circulating neutrophils may be affected by relative exposure to antigen and immune stimulation.

Our vaccine regimen utilizes a replicating virus vector to mucosally prime immune responses followed by systemic immunization to boost them. The response of neutrophils to replicating Ad5hr-recombiniant priming was of particular interest. While the priming immunizations are administered to the upper respiratory tract (Supplementary Fig. 1), the subsequent biodistribution of the vector is broad, and it has been shown to persist for weeks in rectal macrophages (41). Neutrophils are known to provide important immune protection against bacterial and fungal infections, but their role in controlling viral infections has been less well studied (42). They can recognize viruses via pattern recognition receptors, and their responses to viral pathogen associated molecular patterns include respiratory bursts, degranulation, and/or NET formation. Damage associated molecular patterns originating from virally killed cells may also contribute to the activation seen. In this regard, they can have both positive and negative effects against respiratory viral diseases, the latter due to excessive inflammation (43). In HIV infection, some neutrophil responses such as production of ROS, TNFα and IL-8 lead to enhanced viral replication, while NET formation and release of α-defensins have been shown to contribute to viral control (7, 11). Here we evaluated the response of circulating neutrophils to a replicating adenoviral vector. Notably, there were no adverse clinical effects. The neutrophils exhibited increased activation via TNFα (Fig. 2B-C), an increased capacity for B-cell help via IL-21 (Fig. 2F-G), and increased degranulation as seen by increased expression of CD11b and MPO (Fig. 2I, K) on the surface of CD66⁺cells. The frequency of circulating neutrophils decreased following Ad5hr-recombinant administration (Fig. 2A) suggesting a decrease in overall neutrophil numbers. Increased neutrophil apoptosis has been reported during the acute phase of SIV-infection in non-human primates, more severe in pathogenic compared to non-pathogenic models (44). Here, the loss of neutrophils is reflected by the decreased frequencies of CD11b⁺and MPO⁺cells (Fig. 2H, J), despite the increase in the surface expression of these markers and enhanced generation of ROS in response to stimulation (Fig. 5A). Whether the decline in neutrophils following Ad5hr-recombinant administration reflects cell death or migration to mucosal tissue sites targeted by the virus will require further study. Overall, when compared to pre-immunization neutrophils from the same animals, there was a clear improvement in the ability of neutrophils from Ad5hr-recombinant primed animals to mediate not only an innate response, seen in phenotypic and functional enhancement, but also an increased propensity to contribute to B-cell help.

In contrast to the clear response by neutrophils to the mucosally administered replicating Ad5hr-recombinant vaccine, there was no specific response to either ALVAC/Env or DNA&Env following the two systemic boosts. As shown in Fig. 3, a similar pattern of response was observed across all groups including controls and included significant increases in expression of markers related to activation, degranulation, and B-cell help three days after the administration of the second boost. The increases in controls suggested responses were largely due to the adjuvant. Control animals only received alum adjuvant: alum hydroxide for 5 macaques controlling the ALVAC/Env group and alum phosphate for 5 macaques controlling the DNA&Env group (Supplementary Fig. 1). It has been demonstrated in mice that administration of alum quickly results in inflammation at the injection site with increases in inflammatory cytokines and chemokines, recruitment of neutrophils, and formation of NETs leading to neutrophil death (45). The NETosis can lead to accessibility of host DNA which helps drive antigen specific T-cell and B-cell responses. Thus, it is not surprising that initial non-specific responses were observed here 3 days post-systemic boosting in all three macaque groups. The adjuvant effect of alum was nevertheless demonstrated 2 weeks post-boost by development of high-titer SIV Env-specific antibodies and the phagocytic activity of neutrophils, where PMNs from ALVAC/Env- and DNA&Env-vaccinated animals compared to controls mediated significantly better phagocytosis directed against SIV Env-specific antigen, directly correlated with serum Env-specific antibody titers (Fig. 6). There was, however, a difference in the functional ability to generate ROS in response to stimulation between the ALVAC/Env group and the DNA&Env and control groups. There was a degradation of this ability in the ALVAC group two weeks after the second boost, whereas responses in the other two groups were maintained at the post-Ad5hr-recombinant priming levels (Fig. 5B). More investigation is necessary to understand the reasons for this difference. The overall post-boost responses of neutrophils, or PMNs, reflected both non-specific inflammatory responses attributed to alum, as well as the expected immune responses facilitated by adjuvant effects.

Post infection, ongoing viral replication results in ongoing immune stimulation. Nevertheless, neutrophils expressing CD64, TNFα, IL-21, CD11b, and MPO exhibited lower frequencies during acute infection (Fig. 4). These same neutrophil populations exhibited a slight rebound during chronic infection, but mostly maintained lower levels compared to pre-infection timepoints, suggesting a degree of dysfunction. The lower percentages of CD11b⁺and MPO⁺neutrophils, indicating impaired degranulation, were functionally complemented by the lowered ability to generate ROS in response to stimulation by PMNs from chronically infected compared to protected animals (Fig. 5). This has been previously reported by several groups (28, 46, 47). Interestingly, there were positive correlations between MFI levels of IL-21 and TNFα two-weeks post infection and the number of challenges needed for infection, indicating that macaques more resistant to infection maintain neutrophil populations with the propensity for greater B-cell help and activation (Fig. 4N, O).

To directly address the contribution of neutrophils to B-cell help, PBMCs enriched for B-cells were co-cultured with autologous PMNs. The isolated PMNs used were largely comprised of neutrophils, but there is a possibility that the approximate 1.5–5.5% eosinophils and up to 1.5% basophils contributed to the observed effect (48). These estimates are from humans, but in light of these statistics, further experimentation is needed to determine if these other minimally present granulocytes contribute to the effect on B-cells. It has been demonstrated that neutrophils can directly contribute to B-cell help (17), while basophils and eosinophils contribute more indirectly (49, 50). The B-cell help by circulatory neutrophils reported by Puga et al. was only observed after exposure to sinusoidal endothelial cells activated by microbial signals (17), which could be similar to antigenic stimulation via vaccine and adjuvant exposure. Cells considered to be circulatory neutrophils were characterized as CD15 and CD16 high, but as explained above these markers cannot be used for rhesus neutrophils. As the PMNs isolated from rhesus are primarily neutrophils, and in light of the phenotypic and functional data obtained on neutrophils from blood, these granulocytes were used to directly address the contribution of neutrophils to B-cell help.

Rhesus PMNs elicited a significant response from co-cultured autologous B-cells. In comparison to B-cells alone, or B-cells stimulated with CpG, IL-21, and CD40L, B-cells co-cultured with PMNs from protected, infected, and naïve animals had elevated AID expression (Fig. 7), which is essential for class switching and antibody production (51). This effect was further reinforced by the loss of CD27 on B-cells, potentially indicative of blasting and the production of class switched Abs as previously observed in HIV-1 infected individuals (52). There was also a loss of IgD⁺B-cells, particularly in protected and infected animals, and significant activation via Ki67 beyond the level of stimulated B-cells alone. PMNs elicited better B-cell AID expression and a greater loss of IgD, IgM, and CD27 than the stimulation cocktail. Interestingly, B-cells expressing IgM significantly decreased when co-cultured with PMNs (Fig 7J-L), a phenomenon previously observed in mice, where B-cell IgM production decreased in a TGF-β1 dependent manner (32). Taken together, the overall effect of these rhesus PMNs on B-cells is consistent with provision of B-cell help.

In a second co-culture experiment, additional B-cell markers were investigated. In comparison to B-cells only, increases in the frequency of B-cells expressing CD69 and CD80, both indicators of B-cell activation, were observed in the B-cell-PMN co-cultures (Fig. 8A–F), although only B+PMN+IL-10 in naïve animals was significant. Again, in the co-cultures, the percentage of IgD⁺B-cells initially decreased (Fig. 8G-I), whereas at the same time IgG⁺B-cell frequencies increased (Fig. 8J-L), indicating that class switching was occurring. The provision of B-cell help by neutrophils is not well understood mechanistically. When B-cells were separated by a transwell barrier from PMNs, the result was the same as culturing B-cells alone across all three groups of animals and all parameters assayed (Fig. 8). This indicated that this B-cell help was contact dependent. It has been reported that IL-10 induced circulating human neutrophils to provide B-cell help (15). Here, the addition of IL-10 to the PMN-B-cell co-cultures may have had a slight effect, particularly in naïve animals regarding the frequency of CD69⁺B-cells from naïve animals (Fig. 8C). It is possible that other factors not present in our study contributed to previous results. Further elucidation concerning the mechanism of this help and the role of IL-10 will inform future experimentation. Overall, our experiments in macaques demonstrated provision of B-cell help by PMNs, and that this help is largely unaffected by IL-10 and is contact dependent.

The B-cell help observed indicated that class switching and antibody production was stimulated by PMNs, but we further examined blasting and specific antibody production and secretion. Vaccine induced plasmablasts in rhesus macaques have been shown to be CD27^-CD20^-HLA-DR⁺CD38⁺(33). In protected animals, there was a clear increase in these cells when B-cells were co-cultured with PMNs, but not when B-cells were cultured alone or separated from PMNs by a transwell (Fig. 9A), similar to our previous observations. In chronically infected animals (Fig. 9B) there were elevated levels of plasmablast-like cells when B-cells were co-cultured with PMNs, but not when B-cells were cultured alone or separated from PMNs by a transwell. To determine if PMNs would provide B-cell help leading to antibody secreting cells, SIV Env-specific antibody was quantified in culture supernatants from protected and chronically infected animals. Cultures from naïve macaques were not used as they lacked SIV-specific antibody. As expected, B-cells co-cultured with PMNs from both protected (Fig. 9E) and chronically infected animals (Fig. 9G-H) produced more Env-specific IgG and IgA (chronic macaques only) compared to B-cells alone when comparing AUCs, although these differences were not significant. Surprisingly, however, B-cells separated from PMNs in a transwell secreted as much Ig as B-cells and PMNs in contact with one another, clearly distinct from what was observed by flow cytometry (Fig. 8, Fig. 9A-C). We believe this discrepancy can be attributed to the presence of two or more different populations in the enriched B-cells. In the flow cytometry experiments, we examined specific B-cells defined by early markers of B-cell maturation and development up to the point of early plasmablast formation. In contrast, ELISA assays of culture supernatants quantified SIV-specific Ig production by all the B-cells present. Both the protected and chronically infected macaques had been previously vaccinated, and the latter group had also been exposed to continued antigenic stimulation as a result of infection. Thus, both groups had already developed Env-specific memory B-cells as well as plasmablasts and plasma cells (Musich et al., in preparation). We posit that while cell-cell contact might be necessary for provision of help to early B-cells, a soluble factor might be sufficient for stimulating greater antibody production by more differentiated memory B-cells and antibody-secreting cells. Further experimentation will be necessary to determine if this explanation is correct.

Here we have shown that all components of our vaccine regimen, the replicating Ad5hr-recombinants, the adjuvant, and the systemic booster immunizations, have important interactions with neutrophils. Following mucosal priming with the replicating Ad5hr vector, neutrophils displayed phenotypes with a propensity for both greater B-cell help and functionality. The response to the alum was much the same, but at the same time it served as adjuvant to facilitate adaptive immune responses. The high titered antibodies developed in response to the systemic booster immunizations served to mediate potent phagocytic activity by neutrophils. While it has been shown that neutrophils from spleen, and specifically treated circulatory neutrophils, can mediate B-cell help, here we show that antigen and adjuvant exposure can contribute to the ability of neutrophils to mediate B-cell help when cultured together with B-cells. These B-cells are activated, undergo class switching, and produce Abs. This contribution of neutrophils to the adaptive immune response could be harnessed for use in vaccination with the potential to impact protective efficacy of most vaccines currently in use.