Dendritic cell phenotype can be improved by certain chemotherapies and is associated with alterations to p21^waf1/cip1

Abstract

Introduction

Dendritic cells (DCs) possess the capacity to elicit immune responses against harmful antigens and have been used in DC-vaccines to stimulate the immune system to engage cancer cells. However, a lack of an appreciation of the quality of the DC that is used and/or the monocyte from which it is derived has limited their successful incorporation into treatment strategies.

Methods

In the current study, we explored the relationship between cytokine receptor expression on the monocytes and its subsequent development into DCs. The significance of p21 expression in DCs during differentiation was also studied, as was the effect that manipulating this with chemotherapy may have on DC quality.

Results

DCs separated into two groups based on their ability to respond to a maturation stimulus. This quality correlated with a particular receptor profile of granulocyte–macrophage colony-stimulating factor and interleukin 4 expressed on the monocytes from which they were derived. DC quality was also associated with p21 expression, and artificially increasing their levels in DCs by using some chemotherapy improved function.

Conclusions

Overall, these studies have highlighted a role for common chemotherapy in activating p21 in DCs, which is a prerequisite for good DC function.

Introduction

The past few years have seen a rapid increase in the number of therapeutic approaches that attempt to elicit immune responses in patient with cancer, which are tumour specific [1]. Dendritic cell (DC) vaccination is such an approach [2], which is a multistep method that involves the (1) removal of DC precursors from the peripheral blood of patients; (2) induction of a DC-differentiation phase; (3) activation and priming of the DCs in the presence of a tumour antigen, and (4) inoculation of the DC preparation into the patient. The final cellular transfer step can also include the co-administration of adjuvant agents that support the immune-stimulating programme. Clinically, results have not been spectacular, and despite promising results from a number of early phase trials, just one therapeutic cancer vaccine has been approved for use. There are a number of reasons for this disappointing situation and include reduced tumour immune visibility and inappropriately functioning immune effectors [3, 4].

One particular issue has been the definition of an “appropriately matured DC that possesses optimum stimulatory capacity” [5]. The DC-vaccine approach usually involves generating immature DCs ex vivo by culturing peripheral blood monocytes with a standard cocktail of granulocyte–macrophage colony-stimulating factor (GMCSF) and interleukin 4 (IL4). This assumes that monocytes from each patient expresses similar amounts of each receptor for GMCSF and IL4, and equally that they are capable and competent to process these cytokines; however, this is not the case, and expressions do vary between the individuals [6]. Thus, the ability to assess the quality of monocytes before their differentiating into DCs, and/or to pre-screen DC-vaccines prior to transferring back into patients would be clinically advantageous. Similarly, intrinsic heterogeneity between patients mean that some may respond much better to therapeutic vaccination than others, and so a preliminary test to determine the capacity of the immune cells to engage favourably with the therapy could be used to discriminate those most likely to respond. Consequently, patients identified as unlikely to respond could then be offered additional or alternative treatments.

Activation of the receptors for GMCSF and IL4 leads to the activation of a number of signalling cascades, which is an effect that is common to most receptor tyrosine kinases. One element that is triggered by activation of these receptors is the cell cycle inhibitor p21^waf1/cip1 (p21) [7]. This cyclin-dependent kinase has a central role in regulating the cell cycle by guarding the start checkpoint, through which cells need to pass to elicit an S-phase [8]. Although the canonical function of p21 is to control the entry of cells into the cell cycle [9], it can also regulate the activity of NF-κB, STAT3, and Myc that are important entry-point-proteins to intracellular cascades that regulate immune function [10]. Consequently, disruptions to or dysregulations in p21 can impede the normal function of a variety of immune cells [11]. Indeed, p21 levels have been shown to determine T-cell behaviour [12, 13] and also important during the differentiation of myeloid cells [13, 14].

We have recently reported that the quality of DCs can be predicted using the profile of just four genes [5]. The expressions of these genes correlated with the capacity of DCs to mature, and these cells were designated as being “good” or “bad” on the basis of their responses. As a continuation of these research studies, we have examined the receptor profile of monocytes before differentiating them into DCs, and attempted to associate them with the quality of the resultant DC. The hypotheses on which we have been working are that the receptors for GMCSF and IL4 vary in monocytes, and that their relative expressions determine their ability to differentiate into functional DCs. To this end, we have specifically assessed the expression of the receptors in monocytes and correlated expressions with the subsequent ability of their respective DCs to become stimulated by a maturation signal. Furthermore, we have correlated the capacity of DCs to mature with the expressions of p21, and show that chemotherapeutic activation of p21 during the monocyte-to-DC differentiation period is able to improve the quality of DCs that are generated from monocytes.

Materials and methods

Reagents

Artesunate [ART: Pharmacy, St George’s Hospital (SGH), UK] and lenalidomide (LEN: Celgene Corp., Summit, NJ, USA) were dissolved in dimethyl sulphoxide (DMSO), whilst camptothecin (CPT: Sigma Ltd., Dorset, UK), cyclophosphamide (CPM: Sigma), docetaxel (TAX: Sigma), gemcitabine (GEM: SGH), and oxaliplatin (OXP: Sigma) were dissolved in phosphate buffered saline (PBS) to create 10 mM stock solutions that were maintained at −20 °C for no longer than 4 weeks. All controls used in our studies involved treatment with equal amounts of PBS or DMSO (<0.1 %).

Generating immature DCs

Peripheral blood mononuclear cells (PBMCs) were isolated from the blood of healthy donors (National Blood Service, London, UK), and the monocytes harvested by plastic adherence as described previously [5]. An aliquot was removed at this time to assess the expression levels of the receptors for IL4 and GMCSF by incubating the freshly harvested PBMCs with a cocktail of fluorescein isothiocyanate anti-CD14, phycoerythrin (PE) anti-GMCSF-Rα and allophycocyanin (APC) IL4-Rα (all at 1:1,000 and from BD Biosciences, Oxford, UK) for 30 min at 4 °C. Cells were then analysed within 1 h using a FACSCalibur (BD Biosciences). Meanwhile, a DC-differentiating medium [designated DC-diff; basal RPMI-1640 culture medium containing 50 ng/ml IL4 and 100 ng/ml GMCSF (both Peprotech Ltd., London, UK)] was then added to flasks containing the adherent cells (monocytes), which were returned to the incubator for a further 7-days, and fed q.o.d. with DC-diff. The effect of chemotherapy on the differentiation of monocytes to DCs was assessed by supplementing the DC-diff with 100 nM of each of the drugs. After 7-days, non- and loosely adherent cells (DC fraction) were harvested, and the purity assessed by CD11c/HLA-DR/CD14 immuno-discrimination by flow cytometry.

Stimulating DCs: defining responsive and non-responsive cells

Un-stimulated DCs were reset at 1 × 10⁵/ml in basal medium containing a DC-maturing cocktail (DC-mat: 500 U/ml interleukin-6, 500 U/ml interleukin-1β, 500 U/ml tumour necrosis factor-α and 10 ng/ml prostaglandin E) [15] and maintained in a humidified atmosphere with 5 % CO₂ in air at 37 °C for 18 h. The DC-stimulatory effect of chemotherapy was assessed by supplementing the DC-mat with 1 μM of each of the drugs. DCs were then harvested, washed in FACS buffer (PBS containing 1 % (w/v) bovine serum albumin and 0.09 % (v/v) NaN₃), and incubated with a combination of APC anti-CD80 and PE anti-CD86 (both at 1:1,000: BD Biosciences Ltd., Oxford, UK) for 30 min at 4 °C. Acquisition of data was performed within 1 h using a FACSCalibur (BD Biosciences), and the percentage and mean fluorescence intensity of cells expressing the markers determined using the program WinMDI v2.9 (Scripps Research Institute, La Jolla, CA, USA). Furthermore, the ability of the DCs to stimulate the proliferation of allogeneic T-cells was used a simple indicator of DC function, using the methods as previously described [16].

Real time RT-PCR analyses

RNA was extracted from un-stimulated DCs using TRIzol followed by precipitation with iso-propanol. RNA preparation and PCR analysis was performed as described previously [5].

Immunoblotting analysis

Cells were harvested and total cellular protein was solubilised in lysis buffer (New England Biolabs, Hitchin, UK) and resolved by tris–glycine electrophoresis using a 4–12 % bis–tris gradient-gel. Following transfer of proteins to 0.45 μm nitrocellulose membranes, blocking was performed in 5 % (w/v) non-fat milk in TTBS [0.5 % (v/v) Tween-20 in tris buffered saline (Sigma)]. Primary antibody probing was performed with anti-p21 (New England Biolabs and used at a dilution of 1:1,000) and anti-CRBN (1:500; Sigma). Anti-GAPDH was used as a loading control (1:2,000; New England Biolabs), and horseradish peroxidase-conjugated anti-species IgG₁ was used as the secondary antibody (Amersham Biosciences Ltd., Little Chalfont, UK). Bands were visualised by the using SuperSignal West Pico chemiluminescent substrate (Fisher Scientific UK Ltd, Loughborough, UK).

Statistical analysis

All statistical analyses were performed using Microsoft Excel 2003. Any differences between variables and control samples, as determined by analysis of variance, were further characterised by paired analysis tests. Samples were assessed for normality and appropriate parametric or nonparametric tests were used.

Results

DC quality diverges into two groups, which is predictable

Monocytes harvested from healthy blood donors were differentiated into DCs by culturing them in GMCSF and IL4. The purities of the resultant immature DCs were >95 % (mean 98 ± 0.61 %), and these were then matured with a cocktail of inflammatory cytokines [15], which we termed DC-mat. The extent to which DCs displayed CD83 and CD86 was used as an indicator of maturation and/or capacity to mature successfully—that is “how good DCs were in responding to an external signal”. For ease, these DCs were designated “good DCs” [5, 16]. The percentage of DCs expressing high levels of both markers was 87 ± 9.5 % in the cohort of these good DCs; conversely, double-positivity was only 28 ± 1.2 % in bad DCs stimulated with the same cocktail (Fig. 1a, b). The subsequent responsiveness of DCs to DC-mat was predictable, and could be forecast by examining the gene expressions of il8 and cd163 in the immature DC prior to stimulation. The magnitude of gene expression was enumerated by establishing the number of PCR-cycles required to pass a specific level of fluorescence; thus, genes with a low copy number would require more cycles to breach this level. The number of cycles required to meet this level was called the cycle threshold (Ct). Results showed immature DCs that had a relatively low copy number for il8 (as indicated by a higher Ct), but a high copy number for cd163 (low Ct) would predict for a DC that would most likely respond well to stimulation—i.e. be a good DC (Fig. 1a, c).

Fig. 1.

Grouping DCs into categories of good or bad. Freshly differentiated DCs were cultured with DC-mat, and the percentage of cells expressing both CD83 and CD86 were assessed by FACS. Those responding well were designated “Good DCs” and those not called “Bad DCs” (a, b). The expressions of il8 and cd163 in the unmatured DCs were also assessed by real time PCR. Data were presented as a cycle threshold (Ct) which is the number of cycles required for the reporter fluorescence to breach a pre-defined point (see text). These data correlated with their subsequent response to DC-mat (a, c). Specifically, good DCs demonstrated higher Cts for il8, but a lower one for cd163. Ubc was used as the PCR control. Each data point represents the mean of duplicate runs

DC quality correlates with the expression of GMCSF-R and IL4-R on monocytes

DC-diff contains GMCSF and IL4, which are used widely to differentiate monocytes into DCs. Consequently, we surmised expressions of their receptors on monocytes should correlate with their ability to successfully develop into DCs. Monocytes were thus assessed for their expressions of the receptors for these two cytokines (GMCSF-R and IL4-R) using the gating strategy presented in Fig. 2a. To reiterate, we were interested to see if the extent to which these receptors were expressed on monocytes could predict for the quality of DCs that were subsequently generated. Results showed the percentage of monocytes expressing both receptors were the same in both good and bad DCs (49 ± 18 vs. 53 ± 1.9, respectively), and the Pearson’s correlation coefficient (r) was 0.022 (Fig. 2b). There was no link between monocytes that expressed IL4-R but no GMCSF-R and the stimulatory capability of their subsequent DCs (Fig. 2c). However, good and bad DCs could be discriminated on the percentage of monocytes expressing GMCSF-R but not IL4-R (17 ± 2.2 vs. 11 ± 1.8, respectively; p < 0.016). Furthermore, there was good correlation between these monocytes and the quality of the DCs deriving from them (r = −0.817; Fig. 2d).

Fig. 2.

Assessing GMCSF-R and IL4-R on monocytes. The monocytes from which DCs were differentiated were assessed for their levels of the receptors for GMCSF and IL4. The gating strategy is shown in a. Quadrants were then positioned on receptor expressions, and the percentages of monocytes within each determined and correlated with the subsequent ability of their generated DCs to respond to DC-mat (b–d). There was significant negative correlation between the percentages of monocytes expressing just GMCSF-R (GR+ IL4R−) and the percentages of DCs displaying both CD83 and CD86 (d)

p21 expression is associated with DC quality

The induction of p21 has been implicated in DC development as its activation is associated with monocytes differentiating into DCs possessing proficient maturation capacity. For this reason, its level in freshly differentiated but un-stimulated DCs was assessed by immunoblotting, and then correlated with the maturing quality of the respective DCs. Results showed p21 levels were significantly higher in DCs that were successfully responsive to the DC-mat, compared to those that were not. That is, good DCs displayed significantly higher levels of p21 whilst low levels were seen in bad DCs (band density 54 ± 10 vs. 19 ± 8.6, respectively; p = 0.003) (Fig. 3a). Moreover, there was strong positive correlation between p21 expression and DC maturity (r = 0.876). p21 expression on days 1–7 was also assessed in the monocytes as they differentiated into DCs, and their level in good DCs appeared to start low, increase from day 2 and remain elevated on day 7. However, this pattern was inverted in bad DCs, with p21 expression being high on day 1 and low by day 7 (Fig. 3b).

Fig. 3.

p21 expression in DC samples. The amounts of p21 and CRBN were assessed in DCs that had been freshly differentiated from monocytes. The relationship between p21 or CRBN levels and their subsequent responsiveness to DC-mat were compared (a). A fresh set of monocytes were then cultured with DC-diff, and samples removed on days 1, 2, 5, and 7 for analysis of p21 expression; the pattern of which differed depending upon whether they were ultimately good or bad DCs (b). Densitometry results were normalised to GAPDH loading controls

Chemotherapy can increase p21 in DCs during monocytes differentiation

We next tested the idea that increasing p21 levels in fresh DCs could improve their ability to respond to DC-mat. We used common chemotherapy as a way of manipulating/increasing p21, which were incorporated into the monocyte-to-DC differentiation phase. These drugs were used at an equimolar concentration of 100 nM, and administered at the same time as GMCSF and IL4 (q.o.d. for 7-days). We specifically used monocytes that would produce bad DCs to (1) see if their quality could be enhanced by any of the chemotherapy, and (2) see whether any improvements were associated with a greater level of p21.

p21 expression in DCs was initially assessed on day 7, and results showed the drugs impacted them in distinct ways. There was no change to p21 expression following culture with CPM, GEM or OXP; however, the addition of CPT, ART, LEN or TAX to the DC-diff caused significant increases (Fig. 4a). The consequence of increased p21 was then explored by culturing these un-stimulated DCs with DC-mat for 18 h prior to assessing the expression of maturation markers. Results showed that in those DCs where p21 levels had been augmented by chemotherapy, double-positive staining for CD83 and CD86 were increased after stimulation with DC-mat (%double-positive stain for CD83 and CD86 after treatment with CPT, ART, LEN, TAX: 93, 94, 90 and 92 %, respectively, versus 26 % in standard DCs treated with no chemotherapy) (Fig. 4b, c).

Fig. 4.

The effect of chemotherapy on p21 expression. Monocytes were differentiated into DCs using the standard DC-diff before assessing p21 levels on day 7 (Un). In addition to this, 100 nM of different chemotherapy drugs were added to the DC-diff, and p21 levels assessed (a). These freshly differentiated DCs were then matured with DC-mat, and the expressions of CD83 and CD86 assessed by FACS (b); where “Control” refers to the isotype control. There was general concordance between the expressions of p21 and DC maturity (c). These experiments were performed on a monocyte sample that generated bad DCs, which was confirmed by the low %CD83 CD86 in Figure b (“Un”: 25.8 %). DCs were then admixed with CFSE-loaded T-cells at a ratio of 1:10 (DC:T-cell), and the extent of proliferation as indicated by a down-shift in CFSE fluorescence intensity in CD3+ CD4+ cells was assessed (d). Data points in a represent the means and standard deviations (SDs) of at least three separate experiments, and coefficient of variation for each data point was <10 %. SDs were omitted from c for clarity

We then tested the ability of the DCs that had been treated with CPT, TAX, CPM and GEM to induce proliferation in allogeneic T-cells, as a coarse indication of their function. Methodologically, matured DCs were admixed with allogeneic T-cells that had been pre-loaded with CFSE at a ratio of 1:10 (DC:T-cell). Proliferation was then determined by measuring the percentage of CD3+ CD4+ cells with reduced CFSE signals. Results showed that in this model, DCs treated with just DC-mat induced proliferation in 3.4 ± 0.17 % of CD3+ CD4+ cells. This was similar to the level seen with DCs co-treated with CPM or GEM (Fig. 4d). However, DCs that had enhanced CD83 and CD86 expression following co-treatment with CPT or TAX significantly increased the percentage of proliferating T-cells (13.2 ± 1.6 and 12.5 ± 1.5 %, respectively; p < 0.01 when compared to DCs without chemotherapy) (Fig. 4d).

Chemotherapy can improve the action of the DC-mat

Having shown that chemotherapy was capable of influencing the monocyte-to-DC differentiation process resulting in the generation of DCs with improved qualities, we next explored the stimulatory effect that chemotherapy may have directly on DCs that had already been matured. Our idea was that the quality of bad DCs could be improved by using chemotherapy has a stimulating/maturing adjuvant to the standard DC-maturing cocktail. Chemotherapy was thus added to immature DCs for 18 h, which was used either alone or in combination with the DC-mat. The chemotherapy used was selected on the basis of our previous studies that showed that exudates from tumours treated with GEM and OXP were capable of stimulating DCs [5, 16]. The immune-modulatory drug LEN was also included as it is known to increase p21, and is also a drug that we have previously shown to modify DC function.

Results showed that DC-mat alone did not alter the levels of p21 in DCs. Similarly, using GEM or OXP with DC-mat concomitantly also did not increase p21 levels; however, the combination of LEN with DC-mat resulted in an increase in p21 in the DCs (Fig. 5a). These changes to p21 expression were associated with changes to DC maturation. Specifically, chemotherapy alone was incapable of maturing DC, but combining them with DC-mat yielded differences. Specifically, GEM and OXP were incapable of enhancing the stimulatory action of DC-mat as indicated by increases in the percentage of cells expressing both CD83 and CD86 (24 ± 1.2 and 25 ± 2.8 vs. 25 ± 3.1 % where DC-mat was used alone). However, when DC-mat was used concomitantly with LEN, an increase in the percentage of cells expressing both CD83 and CD86 was observed (30 ± 2.5 %) (Fig. 5a). The enhancing effect of LEN was explored further in good and bad DCs, and results indicated that this supporting/improving effect of LEN combined with DC-mat was only seen in the bad DC cohort (Fig. 5b).

Fig. 5.

Chemotherapy enhances the maturing effect of DC-mat. Freshly differentiated DCs were cultured with DC-mat supplemented with 1 μM LEN, GEM or OXP. DCs that were not treated with any chemotherapy were designated NIL. After 18 h, these stimulated DCs were assessed for expressions of p21 and maturation markers. DC-mat significantly matured DCs, whilst the drugs alone were incapable of doing this. There was no further increase in DC maturation when OXP or GEM was combined with DC-mat; however, an increase in %CD83+ CD86+ and a concomitant increase in p21 were seen when cells were co-cultured with DC-mat and LEN (a). This enhancement in DC maturation when using LEN and DC-mat in combination was only seen in bad DCs (b). Data represents the mean and standard deviation of at least three separate experiments

Discussion

This study was initiated in response to our earlier reports defining an ability of exudates from tumour cells to stimulate DC maturation and that some conventional drugs were capable of enhancing this response by modifying the biological make-up of the exudates. In these reports, we showed the quality of these monocyte-derived DCs separated into two groups, which could be forecast by assessing the gene expression profile in the DC prior to stimulation. We argue(d) that this putative form of predictive analysis could prove useful for DC-vaccines, allowing their quality to be confirmed before transfer into patients. Indeed, what is lacking is fundamental quality control testing of monocytes to gauge their differentiation capacity, as those from pathologically abnormal individuals may simply be unsuitable for this ex vivo manipulation. Therefore, the aim of this study was to explore the relationship between the condition of a monocyte and the quality of DCs arising from them. To this end, we assessed on monocytes the levels of the receptors for cytokines required for generating good DCs. Additionally, we explored the role of p21 in this developmental process, and studied the effect that manipulating their expressions with chemotherapy may have on DC quality.

Our earlier reports showed that freshly differentiated DCs exposed to maturation signals would either respond effectively or poorly to a maturation stimulus, and it was on this potential capability to respond that they were defined as a “Bad DC” or a “Good DC”. Therefore, a good DC would respond to a maturation stimulus by significantly increasing DC maturation markers (CD80, CD83, CD86, HLA-DR, CCR7), and equally capable of enhancing the proliferation and growth of T-cells as well as the cytolytic action of CD8+ve T-cells. Similarly, in the presence of a diverse cocktail of common antigens, good DCs were also capable of stimulating autologous T-cells to produce higher levels of IFN-γ. As we showed there was generally good association between increasing CD83 and CD86 and enhanced immune effector function, the markers were used as our primary indicator of DC quality. Our groupings were also supported by gene expression profiling, which was sufficiently distinct for us to distinguish between the two DC groups [5, 17].

These DCs are commonly differentiated from monocytes by using a variety of different cytokine cocktails [18]. Of these, GMCSF and IL4 are the most established combination. The roles of these cytokines in monocyte differentiation are complex, and the quality of DC produced in this manner is dependent upon activation of a number of signalling modules. GMCSF appears to be the principal agent in the mixture and is widely used in combination with other cytokines to differentiate peripheral blood monocytes into DCs [18]. It is unclear what the signal is that determines monocyte differentiation, but the concentration of the cytokine within the microenvironment appears crucial and defines the output of the signalling pathway(s) activated by ligand–receptor binding [19]. Nevertheless, GMCSF action can biased towards the DC route by using IL4, which suppresses the generation of macrophages whilst actively promoting differentiation of monocytes down the DC lineage [20].

Irrespective of the amount of each of GMCSF and IL4, the action of these ligands is consequential only if their cognate receptors are expressed at sufficient levels. For this reason, we assessed the expression of GMCSF-Rα and IL4-Rα on monocytes and compared this with the quality of DCs generated from them. Quality was denoted by the capacity of these DCs to respond to DC-mat. Our results suggested samples consisting of monocytes with a smaller percentage of cells that were simultaneously GMCSF-R positive and IL4-R negative would result in DCs that were more responsive to stimulation. There have been reports of the quality of DCs from monocytes being influenced by the amounts of GMCSF and IL4, or by the availability of their cognate receptors on the monocytes. For example, the extent to which monocyte-derived DCs express, CD1a influences the potency of the signalling through GMCSF, IL4 and their receptors [21, 22]. We and others have thus suggested that it is reasonable to assume that the poor activity of ex vivo-engineered DCs can simply be due to sub-optimal expressions of the cytokine receptors on monocytes [22].

Granulocyte–macrophage colony-stimulating factor and IL4 signal via a number of signalling cascades feed a number of important effectors such as p21 [7]. p21 was of interest and assessed in our studies as it is involved in the process by which monocytes differentiate into DCs [13, 14]. Methodologically, we assessed p21 levels in DC during their differentiation process and also before exposure to DC-mat. Results indicated there was a strong correlation between p21-levels in unstimulated DCs and the responsiveness of the DCs to stimulus (r = 0.876); p21 expression could be used to discriminate between good and bad DCs (p = 0.003). Considering the relationship between p21 and DC quality, we tested the idea that any chemotherapy capable of raising p21 expression may be able to improve the quality of DCs. Our results showed that using ART, CPT, LEN or TAX during the monocyte-to-DC differentiation phase caused an increase in p21 expression on DCs, which resulted in an improvement in the capacity for these DCs to mature. Additionally, these matured DCs were functionally capable of stimulating the proliferation of CD3+ CD4+ T-cells. This was a fascinating observation that suggested that in addition to the conventional action of chemotherapy, modification to p21 levels in the developing DC and subsequent improvements to DC function could also be attributed to them. Parenthetically, although increased p21 in tumour cells have been seen following chemotherapy, and may simply reflect the stressed state of the cell preceding death, we purposely used these drugs at sub-optimal concentrations to ensure DC viabilities, and cell cycle dynamics were not significantly affected (data not shown). Although the agents were used at equimolar dose of 100 nM, it would be interesting to see whether varying the concentrations of each of the drug could influence DC maturation capacity through modifications to p21 levels.

In addition to experiments that explored the effect of chemotherapy on DC-differentiation and p21 expression, we performed experiments that revealed some chemotherapy drugs were capable of increasing p21 levels in DCs post-differentiation. Specifically, in bad DCs, where p21 levels were low, the concomitant use of LEN for 18 h with DC-mat was capable of increasing p21 in these DCs. Furthermore, increased p21 expression was associated with enhanced DC maturation induced by DC-mat. This phenomenon was not replicated by other drugs and suggested that the mechanism by which p21 in DC are induced in this short exposure duration may not be common to all drugs. Indeed, recent reports identifying cereblon (CRBN) as a primary target for lenalidomide encouraged supplementary genomic and proteomic modelling to be performed to survey for putative pathways between CRBN and p21 [23]. The working thought was that p21 was increased by lenalidomide impacting CRBN function. These analyses identified a relationship between them via modifications to the AMP-dependent protein kinase (AMPK) by CRBN. Specifically, CRBN could inhibit AMPK which itself was a modulator of p21 activation [24, 25], and this inverse relationship between CRBN and p21 was confirmed by our Western blotting experiments. Taken together, these data suggest that lenalidomide may have increased the extent to which DCs had been stimulated by modifying the expression of p21 through its interactions with CRBN. We are currently developing new and appropriate experiments to test further the way by which drugs can influence DC functionality. These include assessing the effects that these chemotherapy have on enhancing DC maturation and function in animals carrying syngeneic tumours [26].

The concept that chemotherapy could enhance DC function is at first counterintuitive. However, whereas this would certainly be expected to be the case with high-dose chemotherapies that suppress the bone marrow, it has become apparent that certain chemotherapies can favourably affect the immune response by selectively inhibiting regulatory cells. Low dose cyclophosphamide is used to enhance the immune response by inhibiting T regulatory cells [27], which are much increased in the presence of certain tumour types. It is now clear that this property is not unique to cyclophosphamide. For instance, gemcitabine is very effective at inhibiting myeloid-derived suppressor cells [28], which infiltrate many tumour types such as pancreatic cancer. Whilst exploring this function, we have previously reported that gemcitabine also enhances HLA-1 expression and increases antigen presentation properties [4]. Another way that chemotherapy may enhance DC function is by providing tumour breakdown and antigen release in the right cytokine-enhanced environment.

In conclusion, generating an effective and potent immune response against tumour cells is the fundamental aim of immunotherapy. DC-vaccines have been such an approach that attempts to elicit a tumour-specific response in patients with cancer. Verifying that the cellular material used in the vaccine is immunologically competent prior to their transfer into patients would be a sensible pre-screen test; however, this is not done routinely. The quality of DCs is dependent upon the quality of monocytes from which they are generated, and so, patients with cells that are unsuitable should be tracked into other treatment modalities. Alternatively, adjuvant chemotherapy could be used as a way of improving the differentiation process; an improvement that was connected to changes to p21 levels. Presumably, chemotherapy could be applied ex vivo during DC production, or in vivo; in which case, the conventional effects/benefits of chemotherapy could also be harnessed. It is precisely this combination of immunotherapy with conventional chemotherapy in a way that complements both modalities that is clinically appealing and gaining support.