T helper cell polarisation as a measure of the maturation of the immune response.

BACKGROUND: T helper cell polarisation is important under chronic immune stimulatory conditions and drives the type of the evolving immune response. Mice treated with superantigens in vivo display strong effects on Th subset differentiation. The aim of the study was to detect the intrinsic capacity of T cells to polarise under various ex vivo conditions. METHODS: Purified CD4+ T cells obtained from super-antigen-treated mice were cultured under Th polarising conditions in vitro. By combining intracellular cytokine staining and subsequent flow cytometric analysis with quantitative cytokine measurements in culture supernatants by enzyme-linked immunosorbent assay (ELISA), the differential Th polarising capacity of the treatment can be detected in a qualitative and quantitative manner. RESULTS AND CONCLUSIONS: BALB/c mice were shown to be biased to develop strong Th2 polarised immune responses using Th0 stimulation of purified CD4+ T cells from phosphate-buffered saline-treated mice. Nevertheless, our analysis methodology convincingly showed that even in these mice, Toxic Shock Syndrome Toxin-1 treatment in vivo resulted in a significantly stronger Th1 polarising effect than control treatment. Our results indicate that populations of Th cells can be assessed individually for their differential Th1 or Th2 maturation capacity in vivo by analysing robust in vitro polarisation cultures combined with intracellular cytokine staining and ELISA.


Introduction
Polarisation of T helper cells is important in disease models associated with chronic immune stimulation, where the outcome depends on the type of immune response induced and the relative amount of cytokines released. Many diseases such as leprosy, allergy, multiple sclerosis and responses to immunotoxic agents have pathology associated with aberrant T h 1 and T h 2 polarisation. 1 Á 7 The importance of T-cell polarisation in a disease such as leprosy is immediately obvious since the infection by Mycobacterium leprae can be resolved by a strong cell-mediated T h 1 response, while continued infection is associated with a humoral T h 2 response. 3,5 Gram-positive bacteria including Staphylococcus aureus , Streptococcus pyogenes and Mycoplasma arthritidis secrete a number of exotoxins. These bacteria are well known for their multi-drug resistance in hospitals and are able to induce severe disease. Certain strains of S. aureus produce the superantigen Toxic Shock Syndrome Toxin-1 (TSST-1), which is a pyrogenic toxin and is implicated as a causative factor in toxic shock syndrome. As a superantigen, TSST-1 binds MHC II molecules expressed on antigen-presenting cells (APC) outside the antigenic groove, and stimulate CD4 ' helper T cells (T h ) contingent upon engaging particular T-cell receptor Vb elements. As a result superantigens act as glue, linking T cells of various antigenic specificity to MHC molecules on APC, thereby leading to mutual activation of these cells. In mouse model systems using the Staphylococcal-derived superantigen TSST-1, one of the hallmarks of superantigenic stimulation is the capacity to induce T-cell differentiation towards different T h subsets, including T h 1 and T h 2 type cells. 8 It is hard to determine polarisation of T helper subsets during an acute response since real distinct populations only arise after chronic exposure to a particular antigen. However, many in vivo and in vitro experiments have a limited duration that does not reflect more chronic effects. It is therefore paramount to try to determine in which direction the cell population is polarising. For this reason we performed two secondary ex vivo stimulations under polarising conditions in our system, in order to determine how far our stimulated cells can be pushed in each direction, T h 1 and T h 2. Moreover, unpolarised T h 0 cells can be pushed in any direction in culture by addition of exogenous polarising cytokines in concert with neutralising antibodies to polarising cytokines in the opposite direction, such as interleukin (IL)-4 and interferon (IFN)-g. Resistance to this directional polarisation, or co-expression of IFN-g or IL-4 in the supernatant under opposite polarising conditions, however, indicates a shift towards differentiated T h cell populations. Naturally, these analyses need to be performed on highly purified populations of CD4 ' T cells obtained from circulation or isolated lymphoid organs.
Several techniques can be used to try to determine the direction of polarisation. First, the quantitative reverse transcriptase-polymerase chain reaction can be used to determine the number and nature of cytokine mRNA transcripts of stimulated cells, and can even be performed on a single cell level. 9 This has certain disadvantages, since the actual production and release of cytokines cannot be directly measured and many of the cytokines of interest are subject to post-transcriptional regulatory mechanisms. 10,11 The determination of cytokine production via direct protein detection in supernatant of cultured cells can be done by enzyme-linked immunosorbent assay (ELISA). ELISA determination is a valuable tool to measure the average, overall cytokine response accumulated over time of a heterogeneous population of stimulated cells. However, this method alone is unable to distinguish whether a mixed T h 0 cytokine pattern is due to the release of diverse cytokines from naive T h 0 cells, or rather the production of cytokines by heterogeneous subpopulations of T h 1 and T h 2 cells. Additionally, this method does not allow determination of which proportion of the cells is responsible for this skewing. To work around this limitation and to identify cells producing more than one cytokine, an intracellular cytokine staining (ICS) method is used. This method makes it possible to concurrently check the presence of different cytokines in each cell. 8,9 When combined with ELISA results an impression of the proportion of cytokineproducing cells, as well as the combined cytokine profile of the cell population can be assessed. This approach gives invaluable insight into the direction of T helper cell polarisation.

Splenocyte purification
Mice were sacrificed by cervical dislocation, and spleens removed and placed into single cell suspension, washed then resuspended in 10 ml of RPMI 1640 (Stem Cell Technology, Vancouver, Canada) and centrifuged at 800 g for 7 min. Red blood cells were lysed with Gey's balanced salt lysing solution containing 0.7% NH 4 Cl for removal of red blood cells and washed once in RPMI and then in magnetic activated cell sorter (MACS) buffer.

T-cell polarisation
Ninety-six-well flat-bottomed plates were coated with 50 ml of 10 mg/ml anti-CD3 (clone 145 2C11; Pharmingen, Mississauga, Ontario, Canada) in sterile 0.1 M carbonate buffer (pH 9.6) overnight at 48C. The next day, plates were washed with medium. Subsequently, cells were cultured at 1 )/10 6 cells/ml under the following conditions: T h 0, 2 mg/ml of anti-CD28 (clone 37.51), and 10 U/ml (10 ng/ml) of IL-2 (all Pharmingen); T h 1, as T h 0 plus 10 mg/ml of anti-IL-4, 10 mg/ml of anti-IL-10, 10 ng/ml of recombinant mouse IL-12, and 5 ng/ml of recombinant mouse IFNg (all Pharmingen); and T h 2, as Th0 plus 5 mg/ml of anti-IFN-g, 5 mg/ml of anti-IL-12, and 50 ng/ml of recombinant mouse IL-4 (all Pharmingen). Following 4 days of polarised stimulation, cells were washed twice with cold RPMI. Subsequently, cells were resuspended in fresh complete medium and transferred to a fresh anti-CD3-coated plated prepared as already described, with a final culture volume of 200 ml. Cells were stimulated under T h 0-like conditions in the presence of monensin (Pharmingen) for 5 h if ICS analysis was required or alternatively, cell supernatants were harvested 48 or 96 h after secondary stimulation in the absence of monensin treatment.
FIG. 1. Propidium iodide (PI) staining of stimulated, fixed and permeabilised cells to track cell cycle and to gate out apoptotic cells. PI, and DNA intercalating dye, serves both as a positive control for the permeabilisation of the samples with permwash, but also provides a way of marking cells in the G 0 Á/G 1 , G 2 Á/S and most importantly, apoptotic pathway. Cells with less than one copy of their DNA are undergoing the later stages of apoptosis. Colored gating with the EXPO32 ADC software (Applied Cytometry Systems, Sacremento, California, USA) on the FL3 histogram (B) allows visualisation of the apoptotic cells so that they can be excluded from the live cell gate (A) used in later samples to minimise the fluorescent contamination of staining induced by highly auto-fluorescent dead cells.

ICS staining
Polarised T cells were pelleted in 96-well roundbottomed plates and stained for surface molecules in a 3% foetal calf serum Á/phosphate-buffered saline (PBS) (fluorescent activated cell sorting (FACS)) buffer. Following this, cells were fixed and permeabilised using an intracellular staining kit (Pharmingen) according to the manufacturer's instructions. It's worthwhile to note that fixed, non-permeabilised cells can be stored in FACS buffer at 48C for at least 1 week before ICS is performed. Cells were stained for intracellular cytokines with anti-IL-4-PE (clone 11B11) and anti-IFN-g-FITC (clone XMG1.2) at a 1/ 50 dilution (Pharmingen). Unlabelled antibody of the same clones is utilised at 10 times excess to set negative gates to control for non-specific staining. Cells were resuspended in FACS buffer and then analysed on a BD FACS Calibur (Becton Dickinson Biosciences, Mississauga, Ontario, Canada). A total of 10,000 events in the live cell gate (dead cells excluded by propidium iodine staining) were counted and retained for each sample. Gates were set so that only Â/1% of cells are positive for the cytokine of interest in the labelled/unlabelled negative control, to account for auto-fluorescence and non-specific staining.

Live cell gating
Propidium iodine staining of fixed and permeabilised cells, followed by subsequent gating of cells based on cell cycle and DNA content, provides both a useful live cell inclusion gate for intracellular staining analysis while providing a good positive control for cell permeabilisation during the staining process. Dead cells have obvious changes in forward side scatter, as they are smaller (less FS) and more granular (more SS) (Fig. 1). A total of 10,000 events in the live cell gate are counted and retained for each sample.

ICS-positive gates
The setting of stringent, valid positive gates for ICS is essential, and requires the use of 10 times excess unlabelled anti-cytokine antibody used concurrently with labelled antibody in a known positive control. These criteria are essential, as the use of unstimulated controls alone does not adequately account for changes in autofluorescence seen following activation due to the inevitable increases in forward/side scatter. Due to cell fixation and permeabilisation, there is an inherent amount of non-specific staining that must be accounted for. Typically, the gates are set so that only Â/1% of cells are positive for the cytokine of interest in the labelled/unlabelled negative control (Fig. 2). This background was subtracted from the experimental readings during analysis.

Absolute production
T h 1 polarisation of purified CD4 ' T cells from mice treated in vivo with PBS resulted in significantly higher IFN-g production compared withT h 2 polarised or unpolarised cells (Fig. 3). Production of IFN-g by CD4 ' T cells isolated from mice treated repeatedly with TSST-1 was considerably elevated (n 0/10). Production of IL-4 from polarised CD4 ' T cells was also increased in mice treated in vivo with TSST-1 (Fig. 4). Moreover, T h 1 polarisation resulted in significantly lower IL-4 production than T h 0 or T h 2 polarisation. T h cells from mice treated repeatedly with TSST-1 produce more IFN-g and IL-4 than cells from PBS-treated mice; however, there is more relative polarisation in the T h 2 direction under all stimulatory conditions including T h 0, T h 1 and T h 2.

Relative production
Absolute IFN-g production is made relative by dividing production values by the mean IFN-g concentration following T h 1 polarisation of CD4 ' T cells from in vivo PBS-treated mice, and then expressed as a percentage. Relative T h 2 polarisation is assessed by dividing IL-4 production values by the mean IL-4 concentration following T h 2 polarisation of CD4 ' T cells from in vivo PBS-treated mice, and then expressed as a percentage. Relative cytokine production is assessed by equating cytokine levels detected by ELISA with T-cell polarisation. The average maximal production of IFN-g by CD4 ' T cells isolated from PBS-treated mice stimulated under T h 1 conditions was determined to be the 100% response, and other IFN-g levels were then expressed as a percentage of this maximal production of the positive control. Similarly, the production of IL-4 relative to that by T h 2-stimulated CD4 ' T cells from PBS-treated mice was calculated. Results were graphed to indicate the shifts in the cytokine production of the different T-cell phenotypes (Fig. 5). T h 0 stimulation of CD4 ' T cells from PBS-treated mice clearly shows the BALB/c strains bias towards a strong T h 2 response. T h 1 and T h 2 stimulation of the CD4 ' T cells from the PBS-treated mice demonstrates the relative 100% maximal production of IFN-g and IL-4, respectively. CD4 ' T cells from mice treated repeatedly with TSST-1 have both an enhanced T h 1 and T h 2 response under non-polarising T h 0 stimulation conditions, compared with T h cells from mice treated with PBS. There is a 480% of maximal increase in IL-4 production, while the production of IFN-g is low at 58% of maximal. T h 1 polarisation of the CD4 ' T cells derived from the TSST-1 repeatedly treated mice demonstrates that, despite strong T h 1 polarisation, there is a difference in IL-4 production under this stimulatory condition, indicating that despite strong T h 1 polarisation cells are still actively producing IL-4, indicating a shift to the T h 2 phenotype.

Intracellular cytokine staining
Under non-polarised conditions, ICS staining of CD4 ' cells from the PBS-treated BALB/c mice appeared naturally T h 2 biased, as 5.0% of CD4 ' T cells produced IL-4, while IFN-g production was not detectable (Fig.  6A). However, the BALB/c mice were clearly capable of mounting a strong T h 1 response, as 14.2% of T h 1polarised CD4 ' T-cells were positive for IFN-g (Fig.  6B). T h 2 polarisation of the PBS-treated mice resulted in 4.5% IL-4-positive CD4 ' T cells, a slight decrease from that detected in the T h 0 stimulated group (Fig. 6C).

Discussion
While the conditions in our system clearly induced T h 1 and T h 2 polarised CD4 ' populations, there is room for further optimisation. While we used both IL-12 and IFN-g to induce T h 1 polarisation, it has been reported that IL-12 alone, may be adequate due to its potency to induce IFN-g production in CD4 ' T cells. 10 In addition, the requirement of neutralising antibodies against various polarising cytokines depends on the purity of the CD4 ' population being studied. For example, the addition of anti-IL-12, which is not produced by T cells, but is potentially produced by contaminating monocytes, can be ignored if CD4 ' cell purity is over 90%. In fact, the addition of neutralising antibodies to IL-4 alone has been shown in some systems to be sufficient for inducing T h 1 polarisation. 11 Other researchers commonly use a secondary pulse with PMA 1,12,13 or PHA. 1,13 While this type of secondary stimulation may be less expensive than treatment with anti-CD3, anti-CD28 and IL-2, we feel it is less relevant physiologically. In addition, the method as outlined is easily adaptable to the use of irradiated antigenpulsed APCs in place of the anti-CD3 and anti-CD28 antibodies.
Utilisation of a T-cell polarisation pulse, followed by comparison of relative cytokine production by intracellular cytokine staining under polarised conditions allows ex vivo assessment of the T-cell polarisation state in vivo . Since strong and clear T-cell polarisation typically takes place in chronic disease states or exposure to (potentially toxic) agents, this technique provides an assessment of the current FIG. 5. Relative cytokine production, T h 1 versus T h 2. IFN-g production is made relative by first dividing production values by the mean IFN-g concentration following T h 1 polarisation of CD4 ' T cells from in vivo PBS-treated mice, and then expressing this as a percentage. Relative T h 2 polarisation is assessed by dividing IL-4 production values by the mean IL-4 concentration following T h 2 polarisation of CD4 ' T cells from in vivo PBS-treated mice, and then expressing this as a percentage. T h 0 stimulation of CD4 ' T cells from PBS-treated mice clearly shows the BALB/c strains bias towards a strong T h 2 response. T h 1 and T h 2 stimulation of the CD4 ' T cells from the PBS-treated mice demonstrates the relative 100% maximal production of IFNg and IL-4, respectively. CD4 ' T cells from mice treated repeatedly with TSST-1 have both an enhanced T h 1 and T h 2 response under non-polarising T h 0 stimulation conditions, compared with T h cells from mice treated with PBS. There is a 480% of maximal increase in IL-4 production, while the production of IFN-g is low at 58% of maximal. T h 1 polarisation of the CD4 ' T cells derived from the TSST-1 repeatedly treated mice demonstrates that, despite strong T h 1 polarisation, there is a difference in IL-4 production under this stimulatory condition, indicating that despite strong T h 1 polarisation cells are still actively producing IL-4, indicating a shift to the T h 2 phenotype. direction of polarisation the T cells are heading towards during acute responses, on a population basis via relative cytokine production, and on a cellby-cell basis via ICS. Several caveats must be kept in mind when interpreting data under such conditions, however. Most importantly, the comparison of cytokine production by ELISA must be made on a relative scale, as outlined since the potency of the key T h 1 and T h 2 cytokines, IFN-g and IL-4 respectively, may not necessarily be equal. Similar relative comparisons have been made during micro-array analysis of T h 1 and T h 2 gene transcripts. 14 Additionally, all polarisations must be compared with the results obtained from CD4 ' T cells isolated from a PBS-treated mouse polarised under identical conditions. The rationale for this is apparent when looking at the data demonstrated here, since BALB/c mice are naturally T h 2 biased in their cytokine profile, while other strains, such as SJL mice, are naturally T h 1 biased. 15 The addition of ICS to this method most importantly allows for further assessment of the T helper cell population in experimental groups, to determine whether a mixed cytokine profile detected via ELISA and comparison of relative cytokine production truly represents an undifferentiated T h 0 response of naïve cells, or rather is indicative of a heterogeneous population of simultaneously differentiating T h 1 and T h 2 cells. Additionally, it can allow for further characterisation of the cytokine producing subsets. 12,13 While proportion is the only variable measured here, if this technique is combined with cell surface FIG. 6. IL-4 and IFN-g dual staining of T h 0, T h 1 and T h 2 polarised CD4 ' T cells from repeatedly PBS-treated mice. T h 0 stimulation indicates that 5.0% of the cells were IL-4-positive, while IFN-g-positive cells were of very low number (0.3%). Under T h 1 stimulatory conditions, however, this IL-4-positive population disappears. A total of 14.2% of T h cells under these conditions are IFN-g producers. T h 2 polarisation, however sees a return to low number of IFN-g-positive cells, with IL-4producing cells detected at 4.5%. marker staining (prior to fixation and permeabilisation) or staining for proliferation further important data can be garnered from this robust method.