reliminary studies implicate conventional CD4+ T cells as the sole potential producers of IL-4 following immunisation with antigen preparedn aluminium adjuvants. Furthermore, as GFP positive cells are first detected in the lymph node, our studies indicate that these cells may acto block induction of Th1 responses by aluminium adjuvants. We conclude that devising strategies to block the effects of IL-4 production byhese cells will facilitate the rational design of vaccine adjuvants that induce Th1 responses.
2006 Published by Elsevier Ltd.
eywords: Aluminium hydroxide; Th2; Reporter mice
. Introduction
The induction of a Th2 dominated immune responseemains a major limitation to the application of aluminiumdjuvants to modern vaccines. However, the mechanism(s)hat allow these adjuvants to initiate Th2 responses againstdsorbed antigens remain unclear. IL-4, a major cytokineroduct of Th2 cells has been shown to itself induce Th2esponses in vitro [1,2] suggesting that a primary source ofL-4, other than T cells may act to induce Th2 responses inivo. A number of cell types, such as ��T cells, NK1.1+CD4+cells, NK1.1−CD4+ T cells, eosinophils and cells of theast cell/basophil lineage have the potential to provide a
ource of IL-4 (reviewed [1]). However, the role for an earlyource of IL-4 for induction of Th2 responses in vivo has
∗ Corresponding author. Present address: Centre for Biophotonics, Strath-lyde Institute for Biomedical Sciences, University of Strathclyde, 27 Taylortreet, Glasgow G4 0NR, UK. Tel.: +44 141 548 4022;ax: +44 141 552 4887.
been questioned by studies in mice with disrupted IL-4 pro-duction or IL-4 signaling [3,4]. Similarly IL-4, IL-4R� andStat6 deficient mice immunised with aluminium hydroxide-adsorbed antigens produced normal Th2 responses, com-pared with wild-type mice, strongly suggesting that IL-4is not essential for Th2 induction in vivo [5–7]. However,these studies also identify increased Th1 responses to nor-mally Th2 stimuli in knockout mice, suggesting that themain role of IL-4 in vivo is less to do with the inductionof Th2 type responses and more related to the suppressionof Th1 type responses [6]. Importantly, these studies sug-gest that if IL-4 production induced by aluminium adjuvantscould be antagonised, it may allow Th1 associated cytokinesto dominate and a functional Th1 response to develop [8].To achieve this, it would be helpful to determine whichcells produce IL-4 during an aluminium adjuvant-inducedresponse. This information will allow a better understand-ing of how these adjuvants mediate their effects and facil-itate rational adjuvant design. Here we employ the recentlydescribed, 4get reporter mice [9], that express GFP as part of abicistronic IL-4-IRES-GFP mRNA, to identify IL-4 express-
5394 F. McDonald et al. / Vaccine 24 (2006) 5393–5399
ing cells in situ during an aluminium adjuvant induced Th2responses.
2. Materials and methods
2.1. Mice and inoculations
The generation of 4get mice has been described previ-ously [9] and this strain has been backcrossed to a BALB/cbackground for over 10 generations [10]. BALB/c micewere purchased from Harlan Olac, UK and all animalswere maintained as specified pathogen free under standardconditions in accordance with local and Home Office reg-ulations. Mice were injected subcutaneously in the nape ofthe neck with 100 �l containing 100 �g of ovalbumin gradeVII [Worthington Biochemical Corporation, USA] preparedin 1% alhydrogel [Brenntag Biosector, Frederikssund,Denmark]. This combination of antigen and adjuvant isknown to be endotoxin-free, and fails to stimulate dendriticcells in vitro [11].
2.2. Obtaining cells from lymphoid tissues
Cell suspensions from draining lymph nodes (cervical,brachial, axillary and inguinal nodes) and spleens wereoPSscpfcI
2
Hstc[Aattoto
2
(
mouse serum [Labtech International] and 0.1% azide [Sigma]except for CD49b staining; 20% mouse serum) for 20 minprior to staining. Samples were then incubated at 4 ◦C for20–30 min with monoclonal antibodies against various cellsurface markers. Fluorochrome or biotinylated antibodieswere added to samples and biotin was detected using flu-orochrome labeled streptavidin [Vector, Peterborough, UK].After the primary antibody incubation, samples were washedand resuspended in 200 �l FACS Flow [BD Pharmingen].Some samples were stained with propidium iodide (PI) [BDPharmingen] to allow dead cells to be eliminated from anal-ysis. Antibodies used were specific for CD4-PE, CD4-PerCP(L3T4), B220-PE (RA3-6B2), CD49b-APC (DX5), CD44-PE (IM7), CD25-PE (3C7) and appropriate isotype controlswere run for all samples [all from BD Pharmingen]. All datawas collected by FACScan [BD Pharmingen] and analysedusing FloJo software.
3. Results
3.1. Detection of GFP positive cells in vivo
Analysis of GFP expressing cells was performed on cellsuspensions prepared from draining lymph nodes (Fig. 1a andb), spleens (Fig. 1c and d), injection sites (Fig. 1e and f) andldlilc
3
G3pdqcattaasnsacsstf
btained by forcing tissues through Nitex mesh [Cadischrecision Meshes, London, UK] using a syringe plunger.pleen cells suspensions were resuspended in 1 ml of Boyle’solution [12] and incubated at 4 ◦C for 15 min to lyse redells. Samples were then washed twice in PBS and resus-ended in complete medium (cRPMI; RPMI 1640, 10%oetal calf serum (FCS), 2 mM l-glutamine, 100 U/ml peni-illin, 100 mg/ml streptomycin, 1.25 mg/ml Fungizone [allnvitrogen, Paisley, UK]).
.3. Isolation of cells from peripheral tissues
Tissues from the injection site and lungs were collected inBSS [Invitrogen], forced through nitex and the cell suspen-
ion retained. The aggregated tissue, attached to the nitex washen incubated with 3 ml digestive mix (20% FCS, 210 �g/mlollagenase [Sigma, Dorset, UK] and 30 �g/ml DNAseRoche Diagnostics, Germany] in HBSS) for 20 min at 37 ◦C.fter incubation, the sample was again forced through nitex
nd the resulting suspension centrifuged at 250 × g and addedo the primary cell suspension. This process was repeatedhree times. The resulting cell suspension was then layeredver histopaque [density 1077 g/l, Sigma, UK] (used as perhe manufacturers instructions) and the cells subsequentlybtained were washed in PBS and resuspended in cRPMI.
.4. FACS analysis
All samples were incubated at 4 ◦C with 50 �l Fc blocksupernatant from anti-CD16/CD32 [clone 2.4G2], 10%
ungs (Fig. 1g and h) of mice inoculated with Alum/OVA 11ays earlier. GFP positive cells could be clearly detected inymph node, spleen and lung samples, though their presencen injection sites was more ambiguous (Fig. 1e and f). Inymph nodes, spleen and lung, GFP positive cells were almostompletely CD4+.
.2. Kinetic analysis of GFP expression in vivo
Kinetic analysis of GFP expression demonstrated thatFP+ cells were first detected in draining lymph nodes fromdays after inoculation with Alum/OVA (Fig. 2a). In sam-
les labeled with anti-CD4, all GFP+ cells that could beetected were CD4+. Any GFP+ cells found in the CD4−uadrant fell within the background level of auto-fluorescentells, as shown by the BALB/c control (Fig. 2). Furthermore,ll GFP+ cells were B220− (data not shown). Interestinglyhe frequency of GFP+ cells in the lymph node continuedo rise through the duration of the experiment, suggestingn ongoing process of commitment to IL-4 expression andcquisition of GFP expression. We investigated if it was pos-ible to detect these cells at sites other than the draining lymphodes. Interestingly, GFP+ cells could be identified in thepleen (Fig. 2b) although they were more difficult to reli-bly detect in the spleen than the lymph node and as suchould only be detected with certainty from day 6 onwards,uggesting that the emergence of GFP+CD4+ cells in thepleen occurs later than in the draining lymph nodes. As inhe draining lymph nodes, the general trend in the spleen isor the frequency of GFP+CD4+ cells to increase with time
F. McDonald et al. / Vaccine 24 (2006) 5393–5399 5395
Fig. 1. Analysis of cellular GFP expression in vivo following immunisation with OVA adsorbed to aluminium adjuvant. Cells isolated from draining lymphnodes (a and b), spleen (c and d), injection site (e and f) and lungs (g and h) were assessed by flow cytometry for the presence of CD4+GFP+ or CD4−GFP+cells. Data shows contour plots for control BALB/c (a, c, e and g) and 4get (b, d, f, and h) mice 11 days after immunisation. Data are representative of twoindependent experiments.
5396 F. McDonald et al. / Vaccine 24 (2006) 5393–5399
Fig. 2. The proportion of GFP+CD4+ cells in draining lymph nodes continues to increase until day 11 following immunisation. CD4+ and CD4− cells isolatedfrom draining lymph nodes (a), spleen (b), injection site (c) and lungs (d) were analysed for GFP expression. Data shown for control, BALB/c and 4get micerepresent mean proportion of GFP+ cells within the lymphocyte population ±1S.D. Group sizes for 4get are n = 2 except day 3 and day 8 where n = 4 and forBALB/c are n = 1 except day 3 and day 8 where n = 2. Data are representative of two independent experiments.
and within the limits of the timecourse no evidence for adecrease or a plateau was observed. However, there was apopulation of GFP+CD4− cells present in the spleen thatare not seen in the lymph nodes, or the peripheral tissueswe investigated (Fig. 2c and d). Investigation of the injectionsite revealed that GFP+ cells (CD4+ or CD4−) could notbe identified (Fig. 2c), although generally low cell returnswith low viability, suggested that in situ detection, for exam-ple by immunohistochemistry may be a more appropriatemethod for future analysis. As a control peripheral tissue,that is one that had not been exposed to adjuvant or antigen,we prepared cell suspensions from the lungs. In contrast tothe injection site, GFP+ cells could be detected in the lungs(Fig. 2d) although only at day 11 after immunisation. Again,the GFP+ cell population was almost entirely composed ofCD4+ T cells.
3.3. Analysis of the phenotype of GFP+ CD4 T cells
As well as conventional T cells, some NKT cells, a poten-tial early source of IL-4, are also CD4+. We therefore char-acterised expression of CD4 and GFP within cell popula-tions expressing CD49b, a pan-NK marker. Interestingly, wecould not reliably identify any CD49b+ population in thelymph nodes, although high background staining observedwith the isotype control confounded this analysis (data not
shown). Despite this limitation, CD49b+ populations wereidentified in the spleen and lungs and analysed for expres-sion of CD4 and GFP. In comparison with BALB/c mice,no CD49b+CD4+GFP+ cells were found in either the spleen(Fig. 3a and c) or the lungs (Fig. 3b and d). Interestingly theCD49b+ population isolated from the spleen did contain anotable CD4−GFP+ population.
3.4. Activation status of IL-4 producing cells
The consistent finding of GFP+ cells after day 3, and thecontinued expansion of this population, questioned the acti-vation status of these cells. Thus, we investigated expressionof CD25 and CD44 10 days after inoculation with aluminium-adsorbed OVA, to identify activated and memory T cellsrespectively. In comparison with BALB/c mice, at this time-point the GFP+ population in 4get mice was found to beCD25lo and CD44hi, characteristic of a memory phenotype(Fig. 4).
4. Discussion
In these studies, we have presented early investigationsusing a reporter mouse (4get) to identify the source of IL-4during the Th2 response induced by aluminium adjuvants.
F. McDonald et al. / Vaccine 24 (2006) 5393–5399 5397
Fig. 3. The CD4+CD49b+ population in spleen and lungs are not GFP+. Cells were isolated from tissues of 4get (a and b) and control BALB/c (c and d) miceeight days after immunisation with OVA adsorbed to aluminium adjuvant. Representative FACS plots of PI negative, CD49b positive cells demonstrate that inthe spleen, while a population of CD49b+ cells expressing GFP can be identified (a) these cells are CD4− indicating they are conventional NK cells. Withinthe CD49b+ cell population, no GFP+ cells could be reliably detected in the lungs, irrespective of CD4 status (b). Quadrants for GFP expression were set foreach tissue against the appropriate BALB/c control (shown bottom). Figures in the quadrants show the percentage of the total CD49b+ population residing ineach section. Data are representative of two independent experiments.
While previous studies have demonstrated that IL-4 is dis-pensable for the induction of Th2 responses with aluminiumadjuvants in vivo, they also suggest that IL-4 induced by theseadjuvants actively antagonises the induction of Th1 responses[5,6]. Therefore, identifying the cell population that producesIL-4 in response to these agents would facilitate the ratio-nal modification of this adjuvant for the induction of Th1responses. When trying to identify possible sources of IL-4production in vivo, it is important not only to consider whatcells have the capacity to produce IL-4, but also their anatom-ical location and where they mediate their effects. Thus, as Tcell priming and activation is considered to take place in sec-ondary lymphoid tissues [13] it seems reasonable that theearly source of IL-4, which mediates its effect on preThcells, would be found in these tissues. Our studies demon-strate that in 4get mice immunised with antigen preparedin aluminum adjuvant, the IL-4 reporter signal was entirely
present in CD4+ T cells in draining lymph nodes, through-out the duration of the experiments. While cells of the innateimmune system, in particular eosinophils and cells of themast cell/basophil lineage have been proposed as sources ofIL-4 in infectious disease models [1], the current studies donot support these cells as potential sources of this cytokine indraining lymph nodes after immunisation with antigen pre-pared in aluminium adjuvant. Further studies, have suggesteda subset of CD4+ T cells, NK T cells [14–16] as sources ofIL-4 in response to polyclonal T cell activation. However, inthe current study, the GFP+CD4+ T cell population did notexpress the NK marker, CD49b (DX5), indicating that theywere conventional CD4+ T cells.
Previous studies have demonstrated that Leishmaniamajor infection acts to induce naive T cells to produce IL-4 [17]. It has also been shown that L. major has this effecton antigen-specific responses to heterologous antigens, sug-
5398 F. McDonald et al. / Vaccine 24 (2006) 5393–5399
Fig. 4. Following immunisation with OVA adsorbed to aluminium adjuvant, GFP+ cells assume a memory phenotype. CD4+ draining lymph node cells isolatedfrom control, unimmunised 4get mice (a and c) and 4get mice immunised 10 days earlier were analysed for expression of CD25 (a and b) and CD44 (c and d).Representative FACS plots demonstrate that CD4+GFP+ cells in immunised 4get mice were CD25− (b) and CD44hi (d). Figures represent the percentage ofthe total CD4+ population residing in each gate/quadrant. Data are representative of two independent experiments.
gesting that the parasite has a direct effect on APC popula-tions, rendering them Th2-inducing [17]. Identifying changesin APC phenotypes following administration of aluminiumadjuvants has been, so far unsuccessful [18,19]. However,given that the induction of Th2 responses by these adjuvantsis IL-4 independent, it seems likely that events during antigenpresentation may play a significant role in Th2 polarisation.Nevertheless, in vitro studies as well as in vivo studies usingthe anti-IgD model of Th2 induction, have indicated that IL-4produced by naive CD4+ T cells can act in an autocrine fash-ion, under the direction of APCs, to polarise the developingresponse to Th2 [20,21].
Interestingly, while our results indicate that T cells arethe sole producers of IL-4 in the lymph nodes in responseto aluminium adjuvants, the earliest detection of expressionof the IL-4 reporter in 4get mice was after 3 days post-immunisation, questioning their role as early producers of
IL-4. Furthermore, the acquisition of GFP expression byCD4+ T cells appeared to increase through the course ofthe experiment up to day 11. In contrast, previous studieshave demonstrated that monoclonal TcR transgenic T cellsexpand until day 4/5 and the population declines at or aroundday 7 after immunisation with cognate antigen prepared inaluminium adjuvant [18,22]. This would suggest that onceinitiated, the IL-4 reporter is expressed independently of anti-gen stimulation. Recently it has been shown that while IL-4protein production by differentiating T cells is extremely tran-sient, expression of GFP in 4get cells correlates with thepresence of IL-4 mRNA [23]. Therefore, GFP positive cells,while not necessarily producing IL-4, do reflect cells that arecompetent to produce IL-4 and have an increased capacityto respond to antigen stimulation by further IL-4 production[23]. Dissemination of GFP+ cells around peripheral tissues(the spleen and lung) following immunisation suggest that
F. McDonald et al. / Vaccine 24 (2006) 5393–5399 5399
these cells are either effector or memory T cells. Indeed, thesudden appearance of a large GFP+ population in the lungs atday 11 suggests that these cells are not being induced here, asthe more gradual appearance in the lymph nodes and spleenmight suggest, but that these cells are arriving here from othersites, probably via the blood. At day 10 after immunisation,GFP+ cells in the lymph node displayed a CD25−CD44hiphenotype suggesting that IL-4 competent cells at this timepoint are memory rather than activated, effector cells.
In conclusion, these preliminary results implicate conven-tional CD4+ T cells as the sole producers of IL-4 followingimmunisation with antigen prepared in aluminium adjuvants.This conclusion therefore implicates a central role for APCsin directing the induction of Th1/Th2 responses. GFP positivecells are initially detected in the lymph node. This suggeststhat these cells may act to block induction of Th1 responsesby aluminium adjuvants. Clearly, devising strategies to blockthese effects may be of benefit for the rational design of vac-cine adjuvants to induce Th1 responses.
Acknowledgement
This work was supported by a grant to Dr. J.M. Brewerand Prof. P. Garside from the BBSRC.
R
[7] Cunningham AF, Fallon PG, Khan M, Vacheron S, Acha-Orbea H,MacLennan ICM, et al. Th2 activities induced during virgin T cellpriming in the absence of IL-4, IL-13, and B cells. J Immunol2002;169(6):2900–6.
[8] Brewer JM. How do aluminium adjuvants work? Immunol Lett2006;102(1):10–5.
[9] Mohrs M, Shinkai K, Mohrs K, Locksley RM. Analysis of type2 immunity in vivo with a bicistronic IL-4 reporter. Immunity2001;15(2):303–11.
[10] Gessner A, Mohrs K, Mohrs M. Mast cells, basophils, andeosinophils acquire constitutive IL-4 and IL-13 transcripts duringlineage differentiation that are sufficient for rapid cytokine produc-tion. J Immunol 2005;174(2):1063–72.
[11] Sun H, Pollock KG, Brewer JM. Analysis of the role of vaccineadjuvants in modulating dendritic cell activation and antigen presen-tation in vitro. Vaccine 2003;21(9/10):849–55.
[12] Boyle W. An extension of the 51-Cr release assay for the estimationof mouse cytotoxins. Transplantation 1968;6:761–4.
[13] Mondino A, Khoruts A, Jenkins MK. The anatomy of T-cell acti-vation and tolerance. Proc Natl Acad Sci USA 1996;93(6):2245–52.
[14] Yoshimoto T, Bendelac A, Watson C, Hu-Li J, Paul WE. Role ofNK1.1+ T cells in a TH2 response and in immunoglobulin E pro-duction. Science 1995;270(5243):1845–7.
[15] Yoshimoto T, Paul WE. CD4+NK1.1+ T cells promptly produceinterleukin 4 in response to in vivo challenge with anti-CD3. J ExpMed 1994;179(4):1285–95.
[16] Launois P, Ohteki T, Swihart K, Macdonald HR, Louis JA. In sus-ceptible mice. Leishmania major induce very rapid interleukin-4production by CD4+ T-cells which are NK1.1−. Eur J Immunol1995;25(12):3298–307.
[
[
[
[
[
[
[
eferences
[1] O’Garra A. Cytokines induce the development of functionally het-erogeneous T helper cell subsets. Immunity 1998;8(3):275–83.
[2] Murphy KM, Reiner SL. The lineage decisions of helper T cells.Nat Rev Immunol 2002;2(12):933–44.
[3] Kopf M, Le Gros G, Bachmann M, Lamers MC, Bluethmann H,Kohler G. Disruption of the murine IL-4 gene blocks Th2 cytokineresponses. Nature 1993;362(6417):245–8.
[4] Sacks D, Noben-Trauth N. The immunology of susceptibilityand resistance to Leishmania major in mice. Nat Rev Immunol2002;2(11):845–58.
[5] Brewer JM, Conacher M, Satoskar A, Bluethmann H, Alexan-der J. In Interleukin-4 deficient mice, Alum not only generates Thelper-1 responses equivalent to Freund’s complete adjuvant, butcontinues to induce T helper-2 cytokine production. Eur J Immunol1996;26(9):2062–6.
[6] Brewer JM, Conacher M, Mohrs M, Brombacher F, Alexander J.Aluminium hydroxide adjuvant initiates strong antigen specific Th2responses in the absence of IL-4 or IL-13 mediated signalling. JImmunol 1999;163(12):6448–54.
17] Stetson DB, Mohrs M, Mallet-Designe V, Teyton L, Locksley RM.Rapid expansion and IL-4 expression by Leishmania-specific naivehelper T cells in vivo. Immunity 2002;17(2):191–200.
18] Smith KM, Garside P, McNeil RC, Brewer JM. Analysis of costim-ulatory molecule expression on antigen-specific T and B cells duringthe induction of adjuvant-induced Th1 and Th2 type responses. Vac-cine 2006;24(15):3035–43.
19] Jordan MB, Mills DM, Kappler J, Marrack P, Cambier JC. Promo-tion of B cell immune responses via an alum-induced myeloid cellpopulation. Science 2004;304(5678):1808–10.
20] Noben-Trauth N, Hu-Li J, Paul WE. Conventional naive CD4+ Tcells provide an initial source of IL-4 during Th2 differentiation. JImmunol 2000;165(7):3620–5.
21] Noben-Trauth N, Hu-Li J, Paul WE. IL-4 secreted from individualnaive CD4+ T cells acts in an autocrine manner to induce Th2differentiation. Eur J Immunol 2002;32(5):1428–33.
22] Smith KM, Brewer JM, Rush CM, Riley J, Garside P. In vivo gen-erated Th1 cells can migrate to B cell follicles to support B cellresponses. J Immunol 2004;173(3):1640–6.
23] Mohrs K, Wakil AE, Killeen N, Locksley RM, Mohrs M. A two-stepprocess for cytokine production revealed by IL-4 dual-reporter mice.Immunity 2005;23(4):419–29.
