Nature Immunology
5, 539 - 545 (2004)
Published online: 11 April 2004; | doi:10.1038/ni1062
The serine kinase phosphoinositide-dependent kinase 1 (PDK1) regulates T cell developmentHeather J Hinton1, 2, Dario R Alessi3
& Doreen A Cantrell21 Lymphocyte Activation Laboratory, Cancer Research UK London Research Institute, Lincoln's Inn Fields, London WC2A 3PX, United Kingdom. 2 Division of Cell Biology and Immunology, School of Life Sciences, MSI/WTB Complex, University of Dundee, Dow Street, Dundee DD1 5EH, United Kingdom. 3 MRC Protein Phosphorylation Unit, School of Life Sciences, MSI/WTB Complex, University of Dundee, Dow Street, Dundee DD1 5EH, United Kingdom.
Correspondence should be addressed to Doreen A Cantrell d.a.cantrell@dundee.ac.ukT lymphocyte activation is associated with activation of diverse AGC serine kinases (named after family members protein kinase A, protein kinase G and protein kinase C). It has been difficult to assess the function of these molecules in T cell development with simple gene-deletion strategies because different isoforms of AGC kinases are coexpressed in the thymus and have overlapping, redundant functions. To circumvent these problems, we explored the consequences of genetic manipulation of phosphoinositide-dependent kinase 1 (PDK1), a rate-limiting 'upstream' activator of AGC kinases. Here we analyzed the effect of PDK1 deletion on T lineage development. We also assessed the consequences of reducing PDK1 levels to 10% of normal. Complete PDK1 loss blocked T cell differentiation in the thymus, whereas reduced PDK1 expression allowed T cell differentiation but blocked proliferative expansion. These studies show that AGC family kinases are essential for T cell development.T lymphocyte development in the thymus involves an ordered sequence of differentiation and proliferation and is an essential process for the formation of the peripheral immune system. T cell progenitors entering the thymus lack expression of the major histocompatibility receptors CD4 and CD8 (CD4-CD8- double-negative (DN) cells)1,
2,
3. They initiate rearrangements of the T cell antigen receptor- (TCR) locus (Tcrb) and, if successful, this allows surface expression of a functional pre-TCR complex. The pre-TCR instructs cells to proliferate rapidly and to undergo further differentiation to express CD4 and CD8. CD4+CD8+ double-positive (DP) thymocytes rearrange their TCR locus and express a mature TCR complex and are selected to the CD4+ or CD8+ single-positive (SP) lineage. The key checkpoints of -selection and positive and negative selection are controlled by the pre-TCR and the mature TCR in conjunction with cytokines4. These extracellular stimuli are linked via tyrosine kinases, adaptors and GTPases to a diverse network of intracellular signaling molecules that determine the T cell response and control thymocyte growth and differentiation. Among the molecules linked biochemically to TCRs are serine/threonine kinases, including diacylglycerol-regulated kinases of the protein kinase C (PKC) and protein kinase D (PKD) family and phosphatidyl inositol-3 kinase (PI3K)-controlled serine kinases such as protein kinase B (PKB, also called Akt) and the 70-kDa ribosomal S6 kinase 1 (ref. 5). Genetic studies have shown important functions for tyrosine kinases Lck, Fyn, Zap70 and Syk in thymocyte development6,
7. Similarly, adaptors such as LAT and SLP-76 are required for TCR function, and cells in mice lacking these arrest at the pre-T cell stage and fail to undergo -selection8,
9. In contrast, the function of serine kinases during T cell development is less apparent, although they are known to act downstream of the tyrosine kinases and adaptors that are essential for thymocyte development.
There are considerable biochemical data showing antigen receptor or pre-TCR activation of PI3K- and/or diacylglycerol-dependent signal transduction5,
10. For example, mice lacking expression of the PI3K p110 subunit have a partial defect in T cell differentiation, which is restored after expression of an active PI3K mutant11. The importance of PI3K for T cell development is further indicated by observations that deletion of the phosphatidylinositol(3,4,5) triphosphate (PtdIns(3,4,5)P3) 3-phosphatase PTEN (phosphatase and tensin homolog deleted on chromosome 10), which results in increased PtdIns(3,4,5)P3, causes increases in thymic cellularity12 and can substitute for pre-TCR and interleukin 7 signals and reconstitute thymocyte development in mice deficient in pre-TCR or interleukin 7 signal transduction (M. Naspetti and H. Spits, personal communication). Similar strategies with gain-of-function mutants have also showed that diacylglycerol-regulated kinases such as PKCs and PKD have unique capacities to drive pre-T cell differentiation and allelic exclusion2,
10. Despite the potency of PtdIns(3,4,5)P3 in driving thymocyte proliferation and differentiation (ref. 12 and M. Naspetti and H. Spits, personal communication), deletion of single isoforms of the main targets of PI3K, such as the AGC family serine kinases PKB and p70S6K, has no effect on thymocyte development13,
14. Similarly, deletion of individual isoforms of diacylglyerol-regulated serine kinases such as PKCs affects peripheral T and B cell function but has no discernable consequences on thymocyte development15,
16,
17,
18,
19,
20,
21. One interpretation of these data is that AGC serine kinases are important for mature but not immature T cells. However, serine phosphorylation is one of the most frequent post-translational modifications and it seems unlikely that this major family of serine/threonine kinases has no function in the thymus. A more probable option is that the multiple isoforms of the main AGC kinases in T cells have redundant functions, making it difficult to use the simple genetic strategy of eliminating single molecules to define their function in T cell development. Thus, studies so far may have underestimated the importance of serine kinase networks for T cell development.
One way to overcome potential redundancy in signal transduction molecules is to delete a rate-limiting regulator of key pathways. Many of the serine kinases associated with antigen receptors and cytokines, such as the PKCs, PKB and p70S6K, are members of an evolutionarily conserved AGC kinase family. These AGC kinases all need to be phosphorylated at a 'T-loop' site within their catalytic domain to be activated22. This 'T-loop' phosphorylation is mediated by a 'master kinase', phosphoinositide-dependent kinase l (PDK1), which was first identified in the context of the PI3K-PKB pathway triggered by insulin and other growth factors in fibroblasts and epithelial cells23. Deletion of PDK1 by homologous recombination causes embryonic lethality24. However, to circumvent this prenatal death, mice expressing PDK1 alleles (Pdpk1) flanked with the loxP-Cre excision sequence (Pdpk1fl neo/flneo mice) were used to achieve muscle-specific deletion of PDK1 (ref. 25). These mice are initially viable, although they die early with severe defects in cardiac function. The function of PDK1 in the immune system has not been explored, although T lineage−selective deletion of this kinase would offer a unique opportunity to explore whether AGC kinases are necessary for T cell development in the thymus.
Here we achieved conditional gene deletion of PDK1 in T cells. Complete loss of PDK1 blocked T cell differentiation in the thymus, whereas reduced PDK1 expression allowed T cell differentiation in the thymus to proceed but blocked thymic expansion. PDK1 is thus essential for pre-T cell differentiation, thereby proving unequivocally that AGC serine kinases are essential in thymocyte development.
Results
Deletion of PDK1 prevents T cell development
Pdpk1flneo/flneo mice were bred with transgenic mice expressing Cre recombinase under the control of the proximal p56Lck promoter, which induces expression of Cre in T cell progenitors in the thymus24,
26. We analyzed the resultant LckCre+Pdpk1-/- mice for deletion of floxed Pdpk1 using genomic PCR analysis (Fig. 1a). The LckCre+Pdpk1-/- mice were viable, fertile and healthy; they were of normal size and produced offspring at expected mendelian frequency. However, LckCre+Pdpk1-/- mice had very small thymi and very few thymocytes (Fig. 1b). T cell differentiation can be identified by expression of the coreceptors for the major histocompatibility molecules CD4 and CD8 (ref. 1). Early T cell progenitors are DN cells. After productive rearrangement of the Tcrb locus, the expression of a functional pre-TCR induces a proliferation and differentiation program known as -selection, which is marked by CD4 and CD8 expression1. Thymi from LckCre+Pdpk1-/- mice lacked the normal complement of DP cells and mostly contained DN cells (Fig. 1c). Indeed, DN numbers were modestly increased in LckCre+Pdpk1-/- mice; normal littermate control (NLC) mice (Lckcre-Pdpk1+/+) typically had approximately 2  106 to 5 106 DN cells, whereas LckCre+Pdpk1-/- mice had 5 106 to 9 106 DN cells. In contrast, NLC mice had 100 106 to 200 106 DP cells, whereas in LckCre+Pdpk1-/- mice these numbers were reduced by 99% (1 106 to 2 106).
 | | |
DP thymocytes normally express a mature cell surface TCR and are selected to either CD4 or CD8 SP T cells1. LckCre+Pdpk1-/- mice had very few CD4 and CD8 SP cells (0.1 106 to 0.2 106 versus the NLC complement of 10 106 to15 106; Fig. 1c). Analysis of surface TCR expression showed a very low frequency of cells expressing surface TCR, consistent with the loss of DP cells and SP cells in the LckCre+Pdpk1-/- mice (Fig. 1d). Initial analysis of LckCre+Pdpk1-/- mice thus showed an essential function for this kinase at the DN stage of thymocyte development (Fig. 1c). We consistently found that although the B lymphocyte population was present in the spleen of LckCre+Pdpk1-/- mice, there was a very low frequency of Thy-1+ cells (Fig. 1e). In NLC spleen, Thy-1+ peripheral T cells mainly expressed an TCR and were single positive for CD4 or CD8 (Supplementary Fig. 1 online). The low-frequency Thy-1+ cell subset in the peripheral lymphoid organs of LckCre+Pdpk1-/- mice included mainly CD4SP cells or DN cells and had a very high frequency of  T cells or TCR-deficient cells compared with that of NLC mice. Moreover, the few TCR−positive cells had very low TCR expression (Supplementary Fig. 1 online).
LckCre+Pdpk1-/- thymocytes are blocked at DN4
To further determine the effect of PDK1 loss in the thymus, we studied the development of the earliest T cell progenitors in the DN population. DN T cell precursors can be developmentally identified on the basis of CD44 and CD25 expression: the first progenitors are CD44+ and CD25- (DN1), followed sequentially by the CD44+and CD25+ (DN2), CD44- and CD25+ (DN3) and CD44- and CD25- (DN4) populations1. The proximal p56Lck promoter is expressed in DN1 cells, although its expression can be heterogeneous until the DN3 stage27. We determined the CD25 and CD44 staining profiles of Thy-1-positive CD4 and CD8 DN thymocytes from LckCre+Pdpk1-/- and NLC thymi (Fig. 2a). LckCre+Pdpk1-/- thymi contained both DN3 and DN4 subsets but showed an unusual pattern of CD25 expression on DN3 cells; CD25 was thus higher than normal on DN3 cells (mean fluorescence intensity (MFI), 495, compared with 313 for NLC; Fig. 2a, bottom). CD25 on DN3 cells is controlled by the pre-TCR28. These results are indicative of a problem in pre-TCR function in LckCre+Pdpk1-/- DN3 cells but show that the main block in differentiation in the LckCre+Pdpk1-/- mice occurs in the DN4 subset.
| | |
A principal checkpoint in the thymus is Tcrb locus rearrangement, because only cells that productively rearrange their Tcrb locus express a pre-TCR and are selected to the DN4 stage. To study -selection in LckCre+Pdpk1-/- mice we examined pre-T cells for expression of TCR. Analysis of intracellular TCR expression showed a distinct subpopulation of both NLC and LckCre+Pdpk1-/- DN3 cells that had successfully rearranged their Tcrb locus and accordingly expressed intracellular TCR subunits (Fig. 2b, top). For DN4 cells of NLC mice, 83% expressed TCR subunits intracellularly. These percentages were unchanged in LckCre+Pdpk1-/- mice, consistent with normal selection of Tcrb-rearranged cells from the DN3 to the DN4 stage. It remains to be determined if -selection is normal in terms of allelic exclusion.
One issue was whether LckCre+Pdpk1-/- mice had a global defect in pre-TCR function. To assess this, we examined DN cells for expression of CD2 and CD5, because the expression of both molecules is upregulated by the pre-TCR as cells transit from the DN3 to the DN4 stage29,
30. Flow cytometry showed CD5 expression was upregulated as LckCre+Pdpk1-/- DN3 cells transited to the DN4 stage (MFI, 24−97; Fig. 2b, middle). LckCre+Pdpk1-/- DN4 cells had also upregulated CD2 and indeed had more CD2 than normal (Fig. 2b, bottom). The ability of LckCre+Pdpk1-/- DN4 cells to upregulate CD2 and CD5, which are both pre-TCR-mediated responses10,
31, indicates that pre-T cells from LckCre+Pdpk1-/- mice do not have a global defect in pre-TCR signal transduction.
A principal function of the pre-TCR is to act in synergy with cytokines to regulate cell growth and cell cycle progression of T cell progenitors32. The percentage of LckCre+Pdpk1-/- DN3 cells in the proliferative (S + G2/M) phase of cell cycle was comparable to that in NLC DN3 cells (Fig. 2c, top and middle). DN4 cells from NLC thymi were typically blastoid, highly proliferative cells. In contrast, LckCre+Pdpk1-/- DN4 cells showed a reduced percentage of proliferating cells in the S + G2/M phase of the cell cycle compared with the highly proliferative DN4 cells from NLC mice. Cell size comparison by forward-light-scatter flow cytometry showed that LckCre+Pdpk1-/- DN4 cells were much smaller than NLC DN4 cells (Fig. 2c, bottom). This decease in cell size was the most notable difference between LckCre+Pdpk1-/- and NLC DN4 T cells.
In summary, loss of PDK1 prevented the development of mature T cells. PDK1 loss impaired cell growth and proliferation and inhibited differentiation of -selected T cell progenitors at the DN4 stage of thymocyte development. We considered the possibility that failed production of DP cells was caused by massive apoptosis of thymocytes in LckCre+Pdpk1-/- mice rather than by their failure to differentiate. We directly compared the phenotype of LckCre+Pdpk1-/- thymi with thymi lacking function of the GTPase RhoA, in which failed expansion of pre-T cells is caused by increased apoptosis. In thymi lacking RhoA function, granular apoptotic cells are detected by flow cytometry33,
34. In contrast, there was no phenotypic evidence for increased cell death in thymi; PDK1-null DN4 cells were small but not apoptotic. Moreover, analysis of their cellular DNA content showed normal DNA content compared with that of viable cells, with no evidence for a 'sub-G1' peak of DNA indicative of DNA degradation (Fig. 2c).
Timing of PDK1 loss in LckCre+Pdpk1-/- mice
The first phenotypic difference between LckCre+Pdpk1-/- and NLC thymi was seen at the DN3 stage, at which LckCre+Pdpk1-/- cells expressed more CD25 than NLC cells, yet seemed to proliferate and transit to the DN4 stage. LckCre+Pdpk1-/- DN4 cells were more obviously abnormal, as they were very small, failed to enter cell cycle properly and moreover failed to upregulate CD4 and CD8 and transit to the DP stage. To fully understand these data, we studied the timing of loss of PDK1 function in LckCre+Pdpk1-/- mice. The Lck promoter used to drive expression of the Cre transgene should express at the DN1-DN2 stage. However, genomic PCR analysis for deletion of floxed Pdpk1 indicated deletion was partial in DN3 cells and was complete in DN4 cells (Fig. 1a). Accordingly, LckCre+Pdpk1-/- DN3 cells should have reduced PDK1, whereas DN4 cells should have no PDK1.
The deletion of Pdpk1 would not necessarily produce immediate loss of PDK1 protein or function. Therefore, we sought to establish at what stage LckCre+Pdpk1-/- mice lost functional PDK1. One substrate for PDK1 is p70S6K, which has a dual requirement for PDK1: it needs to be phosphorylated by PDK1 at its 'T-loop' site to be activated35. In addition, p70S6K activation is dependent on PKB, another PDK1 substrate that regulates p70S6K through modulation of tuberous sclerosis complex-1/2 function36. The intracellular activity of p70S6K can be monitored by quantification of phosphorylation of its substrate, the ribosomal S6 subunit, by intracellular staining with a specific phosphorylated S6 (phospho-S6) antisera. This is a sensitive assay for PDK1 function because it analyzes phosphorylation of a downstream target of the enzyme. Moreover, the specificity of the assay can be readily verified because treatment of cells with rapamycin, which inhibits the activity of mammalian target of rapamycin (mTor), rapidly reverses S6 phosphorylation36,
37. Phospho-S6 staining of rapamycin-treated cells thus provides an internal negative control as a standard for each sample. Ex vivo wild-type DN3 cells and DN4 cells had high basal phospho-S6 that was lost when cells were pretreated with rapamycin (Fig. 3a). Hence, DN3 and DN4 T cell progenitors activate PDK1 signaling while in situ in the thymus. LckCre+Pdpk1-/- DN4 cells had no basal phospho-S6 (Fig. 3b). In contrast, LckCre+Pdpk1-/- DN3 cells showed residual S6 phosphorylation, although there was a subset of cells in which specific S6 phosphorylation was reduced to undetectable amounts (overlay with negative control; Fig. 3b). The fact that LckCre+Pdpk1-/- DN3 cells can progress to the DN4 stage must be considered with the fact that DN3 cells have initiated but not completed PDK1 loss.
| | |
Reduced PDK1 expression regulates T cell development
Although LckCre+Pdpk1-/- DN3 cells had some residual PDK1 signaling, PDK1 loss must have some effect, as shown by the higher-than-normal CD25 expression in these cells, a phenotype associated with defective pre-TCR function. These results suggest that thymocytes might be sensitive to reduced PDK1 expression. Mice with hypomorphic Pdpk1 alleles with reduced expression of PDK1 have been generated (Pdpk1-/fl)24. These mice are viable, albeit smaller than normal (by approximately 30%), and quantification of PDK1 indicates a general reduction in PDK1 to approximately 10% of normal in a wide range of different tissues, including lymphoid organs24. To probe further the function of PDK1 in the immune system, we analyzed the T cell status of mice with hypomorphic Pdpk1 alleles. Thymocyte numbers in Pdpk1-/fl hypomorphic mice were slightly lower than normal but were within the normal range, given the 30% overall reduction in size of the hypomorphic mice (Fig. 4a). However, analysis of the spleen showed an 80% reduction in peripheral T cell numbers, which is more severe than the proportionately predicted 30% reduction in peripheral T cell numbers (Fig. 4b). Analysis of thymi of Pdpk1-/fl hypomorphic mice showed normal ratios of DN, DP and SP cells (Fig. 4c, top). A closer analysis of DN cells showed a subtle but reproducible effect of reduced PDK1 expression: all four DN subsets were present, but CD25 expression on CD25+ cells was higher than normal in the Pdpk1-/fl hypomorphic mice (MFI, 457, compared with 223 for NLC). This pre-T cell phenotype was reminiscent of that seen in LckCre+Pdpk1-/- thymi (Fig. 4c, bottom and Fig. 2a, bottom, respectively).
| | |
The reduction in peripheral T cell numbers in Pdpk1-/fl hypomorphic mice could be due to their development in an impaired thymic environment, rather than being due to a T cell−autonomous effect or defect. To examine this, we assessed the ability of Pdpk1-/fl hypomorphic bone marrow to competitively repopulate T cells in lethally irradiated wild-type mice. In these experiments, lethally irradiated mice were reconstituted with bone marrow from Pdpk1-/fl hypomorphic or wild-type mice mixed in a 1:1 ratio with bone marrow from wild-type mice. The allelic markers CD45.2 and CD45.1, respectively, were used to distinguish the two donor populations; cells derived from wild-type CD45.1+ bone marrow served as an internal control. We analyzed these mixed bone marrow chimeras over subsequent weeks for population of the thymus with wild-type or Pdpk1-/fl hypomorphic T cells. At day 8 after reconstitution, small numbers of donor T cell precursors could be detected within the CD4 and CD8 DN compartment of the thymi of reconstituted mice (Fig. 5a). These were mainly DN3 cells (data not shown), irrespective of whether donor thymocytes were derived from wild-type or Pdpk1-/fl hypomorphic thymi. By day 10, DP donor thymocytes were present, and in terms of the ratio of DN and DP subsets there was no discernable difference between wild-type and Pdpk1-/fl hypomorphic thymocytes (Fig. 5b). At these time points, the numbers of donor-derived thymocytes were very low but similar, irrespective of whether the thymocytes were derived from wild-type or Pdpk1-/fl hypomorphic bone marrow, indicating that Pdpk1-/fl hypomorphic bone marrow can 'seed' the thymus with T cell progenitors. We determined the numbers of donor-derived thymocytes in the bone marrow reconstituted mice at different days after reconstitution (Fig. 5c). There was notable proliferative expansion of thymocyte populations derived from wild-type bone marrow, but this expansion was absent from the thymocyte populations derived from Pdpk1-/fl hypomorphic bone marrow (Fig. 5c). Analysis of mixed bone marrow chimeras 4−8 weeks after reconstitution showed very few Pdpk1-/fl hypomorphic T cells (Fig. 5d). Thus, Pdpk1-/fl hypomorphic bone marrow cannot compete with wild-type bone marrow to reconstitute to the T lymphocyte compartment in mixed bone marrow chimeras. In contrast, Pdpk1-/fl hypomorphic bone marrow did make a substantial contribution to the B cell compartment; there was a slight reduction in Pdpk1-/fl hypomorphic B cells relative to wild-type B cells, but this was a mild effect compared with the T cell phenotype (Fig. 5d). Despite the small numbers of Pdpk1-/fl hypomorphic thymocytes, analysis of the thymus showed reduced PDK1 expression was not associated with a block in differentiation (Fig. 5e). The main thymocyte subsets, CD4 and CD8 DN cells, DP cells and SP cells, were present in Pdpk1-/fl hypomorphic mice, although they were considerably reduced in number. Similarly, although there were reduced numbers of peripheral T cells in Pdpk1-/fl hypomorphic mice, they had the normal ratio of CD4 and CD8 SP cells. These data are consistent with normal differentiation but reduced proliferation in Pdpk1-/fl hypomorphic thymocytes.
| | |
Discussion
Many studies have suggested that AGC family serine kinases such as PKCs, PKB or p70S6K are essential in peripheral T cell growth and differentiation21,
38. Our results now show the effect of loss of AGC serine kinase function for T cell development and show that that PDK1, a 'master kinase' upstream of diverse AGC kinases, has a key function in the thymus. Pre-T cells activated PDK1-mediated signal transduction pathways in vivo. Moreover, experiments using the Cre recombinase under control of the Lck promoter to conditionally delete PDK1 alleles in pre-T cells showed that PDK1 was essential for T cell development. LckCre+Pdpk1-/- mice thus had very small thymi and very low frequencies of Thy-1+ T cells in their peripheral lymphoid organs. Moreover, the few Thy-1+ cells present were not phenotypically normal TCR−high cells but had a high percentage of T cells and TCR-null cells were found. The presence of T cells in LckCre+Pdpk1-/- mice probably reflects the fact that the Lck promoter−driven Cre recombinase deletes PDK1 after lineage commitment39. The failure to produce mature T cells could occur because of failed thymocyte differentiation or proliferation. The main reason for failed T cell development in LckCre+Pdpk1-/- mice was that most LckCre+Pdpk1-/- DN4 cells failed to upregulate CD4 and CD8 expression and failed to differentiate to the DP stage. Moreover, LckCre+Pdpk1-/- DN4 cells were smaller than normal, indicating that PDK1 is essential for optimal cell growth in this thymocyte subset. There was no evidence that PDK1 loss caused any increase in apoptosis of DN4 cells, but there was a reduction in the frequency of cells in the proliferative S and G2 phases of the cell cycle for LckCre+Pdpk1-/- DN4 cells. Defective proliferation would not by itself result in a block of thymocyte differentiation, as there are many examples of genetically modified mice with very small thymi that contain the normal DN, DP and SP subsets; for example, mice lacking the common cytokine receptor -chain (M. Naspeti and H. Spits, personal communication). It would thus seem that PDK1 is needed to control the transcriptional responses essential for T cell differentiation from DN cells to the DP stage and is also involved in regulating pre-T cell proliferation and growth. One explanation for the smaller size of LckCre+Pdpk1-/- DN4 cells was reduced protein synthesis, as these cells had no active p70S6K, a molecule with an evolutionarily conserved, pivotal function in the regulation of cell growth40.
The signal transduction pathways that control thymocytes are often recapitulated in mature T cells. Accordingly, the effect of PDK1 deletion on T cell differentiation in the thymus suggests that inhibitors of this molecule would have a substantial effect on the adaptive immune system. PDK1 could thus be an important target for immunomodulatory drugs. However, gene-deletion studies have limited use as predictors of in vivo consequences of pharmacological inhibitors because most inhibitors will reduce rather than totally abrogate enzyme function. It is thus more relevant to probe T cell function in mice with reduced expression of PDK1, which have more potential to model the in vivo effect of a pharmacological inhibitor of PDK1. Here we examined the effect of a reduction in PDK1 expression by assessing T cell development in Pdpk1-/fl hypomorphic mice, in which PDK1 expression is 10% of normal. Notably, Pdpk1-/fl hypomorphic mice made few mature peripheral T cells. However, reduction of PDK1 expression was not associated with a block in thymocyte differentiation: the main thymocyte subsets, DN cells, DP cells and SP cells, were present at the normal ratios in thymi of Pdpk1-/fl hypomorphic mice. This suggests that reduction in PDK1 expression inhibits thymocyte proliferation but not differentiation. We tested this hypothesis using bone marrow reconstitution experiments, which confirmed that T cells differentiate with reduced PDK1 expression but fail to proliferate. These experiments also showed that these defects were cell autonomous and were not due to any defects in thymic stromal cells. A major point of thymocyte population expansion occurs as -selected DN4 cells transit to the DP stage1. Analysis of the kinetics of the differentiation of Pdpk1-/fl hypomorphic T cell progenitors in the competitive bone marrow reconstitution experiments showed these populations did not undergo the normal proliferative expansion that occurs at this transition point. Loss of PDK1 or a reduction in PDK1 expression in T cell progenitors had a similar outcome: defective production of peripheral T cells. However, the underlying basis for the failed T cell development in these two conditions is different. Thus, complete loss of PDK1 blocks T cell differentiation, whereas reduced PDK1 expression affects T cell proliferation. That PDK1 deletion and reduction have different consequences shows that PDK1 substrates must differ in the quantity of PDK1 they require for activation. The ability of quantitative differences in PDK1 to yield qualitative differences in T cell responses also indicates that this molecule can translate signal strength into diverse responses.
T cells seem to be very sensitive to decreases in PDK1 expression compared with B lymphocytes. Pdk1-/fl hypomorphic bone marrow can reconstitute the B cell pool with only slightly reduced efficiency, compared with its notable loss of ability to reconstitute the T cell pool. An explanation for this discrepancy could be that PDK1 expression is not equally reduced in Pdpk1-/fl hypomorphic T and B cell progenitors. PDK1 links the actions of PI3K to networks of serine kinases, and studies on the effect of gene deletion of different PI3K subunits have given the impression that B cell development is more dependent on PI3K activity than T cell development41,
42. This discrepancy could be because T cells express multiple isoforms of PI3K, which may have compensating functions, whereas there is no kinase to compensate for the loss of PDK1. The apparent insensitivity of B cells to reduced PDK1 expression could reflect the fact that other mediators of PI3K actions such as Tec family tyrosine kinases or Rac and Rho GTPases are more important for B cells than are PDK1 and its target kinases43.
In summary, our study has shown that PDK1 is essential for T cell development and hence establishes a key function for serine kinases in thymocyte development. Previous studies have established the importance of PDK1 in insulin signaling and in maintaining cardiac function24,
25. It has also been proposed that PDK1 might be a valuable drug target for the treatment of cancer in conditions in which there is deregulation of PI3K signaling pathways44; for example, in transformed cells lacking expression of the PtdIns(3,4,5)P3 3-phosphatase PTEN. The extreme sensitivity of T cells to decreases in PDK1 expression indicates that pharmacological manipulation of PDK1 activity would have a considerable effect on the adaptive immune system. This might be counterproductive for tumor therapy and indicates that PDK1 might be a more valuable target for immunotherapy of T cell−mediated diseases.
Methods
Mice.
Pdpk1flneo/flneo and Pdpk1-/fl hypomorphic mice have been described24. Pdk1-/fl hypomorphic mice have one allele in which exons 3 and 4 have been removed by Cre recombinase, resulting in a frameshift mutation that ablates expression beyond exon 2 (ref. 45). The other Pdpk1 allele contains a neomycin-resistance gene flanked by two loxP sites between exons 2 and 3 as well as a loxP site after exon 4 (Supplementary Fig. 2 online). This allele is not expressed effectively, and Pdpk1-/fl hypomorphic mice have PDK1 expression reduced to 10% of normal in a wide range of tissues24,
45. In Pdpk1flneo/flneo mice, the neomycin-resistance gene within the Pdpk1 alleles has been excised by Cre recombinase; both Pdpk1 alleles have a loxP site flanking exons 3 and 4. LckCre+Pdpk1-/- mice were generated by crossing Pdpk1flneo/flneo mice with transgenic mice expressing Cre recombinase under control of the proximal p56Lck promoter26 (Supplementary Fig. 2 online). This cross results in a T cell−specific deletion of PDK1. Mice were bred and maintained in pathogen-free conditions at Cancer Research UK Biological Resources Unit. Animal experimentation was approved by the Cancer Research UK Research Services Animal Ethics Committee and was done under UK Home Office project licenses PPL60/3116 and PPL70/4458.
Bone marrow chimeras were generated by intravenous injection of bone marrow cells into lethally irradiated recipients carrying the allele encoding CD45.1 on a C57/B6 background (Ly5.1; Charles River Laboratories). Recipient mice were rescued with bone marrow from Pdpk1-/fl (CD45.2)/wild-type (CD45.1) or wild-type (CD45.2)/wild-type (CD45.1) mice at a ratio of 1:1. Chimeras were analyzed 8 days to 8 weeks after the bone marrow transplant.
Flow cytometry and cell sorting.
Antibodies conjugated to fluorescein isothiocyanate (FITC), phycoerythrin, allophoycocyanin and biotin were obtained from PharMingen. TriColor and allophoycocyanin-indotricarbocyanine−conjugated antibodies were obtained from Caltag. Cells were stained for surface expression of the following markers using the antibodies in parentheses: CD4 (RM4-5), CD8 (53-6.7), CD25 (7D4), CD44 (IM7), CD5 (53-7.3), CD2 (RM2-5), Thy-1.2 (53-2.1), TCR (H57-597), CD3 (145-2C11), B220 (RA3-6B2), Gr1 (RB6-8C5), CD11b (M1/70) TCR (GL3), pan-NK (DX5), CD45.1 (A20) and CD45.2 (104). Cells were stained with saturating concentrations of antibody and data were acquired on a FACSCalibur or a BD LSRI (Becton Dickinson). Events were collected and stored ungated with CellQuest (Becton Dickinson) software. Data were analyzed with CellQuest or FlowJo (Treestar) software. Live cells were gated according to their forward-scatter and side-scatter profiles.
CD4 and CD8 DN subsets were gated by lineage exclusion of all CD4, CD8 DP and SP cells, as well as cells of non-T cell lineages, with a panel of biotinylated antibodies (to CD4, CD8, CD3, TCR, CD11b, Gr1, B220 and NK)10. Mature SP thymocytes were defined as Thy-1+, TCRhi and SP for CD4 or CD8. For purified cell populations, cells were stained for the relevant surface markers and sorted with a MoFlo cell sorter (Dako Cytomation). The sorted populations were reanalyzed by FACSCalibur and were found to be more than 95% pure.
Cell cycle analysis.
The cellular DNA content of DN3 and DN4 thymocytes was analyzed on live, saponin-permeablized cells with 7-amino actinomycin D staining. Thymocytes (5 106) were surface stained for CD25, Thy-1 and lineage markers (including CD44). Cells were incubated for 1 h at 37 °C in 800  l 7-amino actinomycin D staining solution (20 g/ml 7-amino actinomycin D, 25 g/ml RNase and 0.03% saponin in PBS supplemented with 20 mM HEPES, pH 7.4, and 2% heat-inactivated FBS). Events were collected on a FACSCalibur with CellQuest software and analyzed with a doublet discrimination module.
Intracellular TCR staining.
Thymocytes were stained for cell surface markers to define DN3 and DN4 subsets. After fixation in 1% paraformaldehyde for 10 min at 25 °C, cells were washed in PBS and permeabilized for 10 min at room temperature in saponin buffer (0.5% (weight/volume) saponin, 5% FBS and 10 mM HEPES, pH 7.4, in PBS) Permeabilized cells were incubated for 45 min with phycoerythrin-conjugated antibody to TCR in saponin buffer, were washed in saponin buffer and were analyzed on a FACSCalibur. Cell surface binding sites were blocked by biotinylated TCR and the specificity of staining was controlled by parallel staining with phycoerythrin-conjugated isotype-matched control antibody (Armenian hamster IgG2 ).
Intracellular phospho-S6 staining.
Thymocytes were treated with 20 nM rapamycin or were left untreated for 20 min at 37 °C. Treatment of cells with rapamycin inhibits the activity of mTor and rapidly reverses S6 phosphorylation36,
37. Phospho-S6 staining of rapamycin-treated cells thus provides an internal negative control as a standard for each sample. Cells were washed and stained with surface markers to define the DN3 and DN4 subsets, then were fixed in 0.5% paraformaldehyde for 15 min at 37 °C, followed by 15 min in 90% methanol on ice. After fixation, cells were washed twice in BSA buffer (0.5% BSA in PBS), then blocked for 10 min at 25 °C in BSA buffer. Cells were incubated for 30 min at 25 °C with antibody to phospho-S6 (2211; Cell Signaling Technologies) in BSA buffer, then were washed and incubated for 30 min at 25 °C with FITC-conjugated donkey IgG antibody to rabbit (Jackson ImmunoResearch). Samples were washed in BSA buffer and were analyzed on a FACSCalibur.
Note: Supplementary information is available on the Nature Immunology website.
Received 1 December 2003; Accepted 18 February 2004; Published online: 11 April 2004.
REFERENCES
-
Borowski, C. et al. On the brink of becoming a T cell. Curr. Opin. Immunol. 14, 200–206 (2002). | Article | PubMed | ISI | ChemPort |
-
Michie, A.M., Soh, J.W., Hawley, R.G., Weinstein, I.B. & Zuniga-Pflucker, J.C. Allelic exclusion and differentiation by protein kinase C-mediated signals in immature thymocytes. Proc. Natl. Acad. Sci. USA 98, 609–614 (2001). | Article | PubMed | ChemPort |
-
Michie, A.M. & Zuniga-Pflucker, J.C. Regulation of thymocyte differentiation: pre-TCR signals and -selection. Semin. Immunol. 14, 311–323 (2002). | Article | PubMed | ISI | ChemPort |
-
Spits, H. Development of T cells in the human thymus. Nat. Rev. Immunol. 2, 760–772 (2002). | Article | PubMed | ISI | ChemPort |
-
Cantrell, D.A. Regulation and function of serine kinase networks in lymphocytes. Curr. Opin. Immunol. 15, 294–298 (2003). | Article | PubMed | ISI | ChemPort |
-
Acuto, O. & Cantrell, D. T cell activation and the cytoskeleton. Annu. Rev. Immunol. 18, 165–184 (2000). | Article | PubMed | ISI | ChemPort |
-
Cantrell, D.A. Transgenic analysis of thymocyte signal transduction. Nat. Rev. Immunol. 2, 20–27 (2002). | Article | PubMed | ISI | ChemPort |
-
Samelson, L.E. Signal transduction mediated by the T cell antigen receptor: the role of adapter proteins. Annu. Rev. Immunol. 20, 371–394 (2002). | Article | PubMed | ISI | ChemPort |
-
Koretzky, G.A. et al. Regulation of hematopoietic cell development and activation by adapter proteins. Immunol. Res. 27, 357–366 (2003). | Article | PubMed | ISI | ChemPort |
-
Marklund, U., Lightfoot, K. & Cantrell, D. Intracellular location and cell context-dependent function of protein kinase D. Immunity 19, 491–501 (2003). | Article | PubMed | ISI | ChemPort |
-
Rodriguez-Borlado, L. et al. Phosphatidylinositol 3-kinase regulates the CD4/CD8 T cell differentiation ratio. J. Immunol. 170, 4475–4482 (2003). | PubMed | ISI | ChemPort |
-
Suzuki, A. et al. T cell-specific loss of Pten leads to defects in central and peripheral tolerance. Immunity 14, 523–534 (2001). | Article | PubMed | ISI | ChemPort |
-
Shima, H. et al. Disruption of the p70s6k/p85s6k gene reveals a small mouse phenotype and a new functional S6 kinase. EMBO J. 17, 6649–6659 (1998). | Article | PubMed | ISI | ChemPort |
-
Peng, X.D. et al. Dwarfism, impaired skin development, skeletal muscle atrophy, delayed bone development, and impeded adipogenesis in mice lacking Akt1 and Akt2. Genes. Dev. 17, 1352–1365 (2003). | Article | PubMed | ISI | ChemPort |
-
Leitges, M. et al. Targeted disruption of the
 PKC gene results in the impairment of the NF- B pathway. Mol. Cell. 8, 771–780 (2001). | Article | PubMed | ISI | ChemPort |
-
Leitges, M. et al. Protein kinase C- is a negative regulator of antigen-induced mast cell degranulation. Mol. Cell. Biol. 22, 3970–3980 (2002). | Article | PubMed | ISI | ChemPort |
-
Leitges, M. et al. Immunodeficiency in protein kinase c-deficient mice. Science 273, 788–791 (1996). | PubMed | ISI | ChemPort |
-
Castrillo, A. et al. Protein kinase C is required for macrophage activation and defense against bacterial infection. J. Exp. Med. 194, 1231–1242 (2001). | Article | PubMed | ISI | ChemPort |
-
Mecklenbrauker, I., Saijo, K., Zheng, N.Y., Leitges, M. & Tarakhovsky, A. Protein kinase C controls self-antigen-induced B-cell tolerance. Nature 416, 860–865 (2002). | Article | PubMed | ISI | ChemPort |
-
Su, T.T. et al. PKC- controls IB kinase lipid raft recruitment and activation in response to BCR signaling. Nat. Immunol. 3, 780–786 (2002). | Article | PubMed | ISI | ChemPort |
-
Sun, Z. et al. PKC-
 is required for TCR-induced NF-B activation in mature but not immature T lymphocytes. Nature 404, 402–407 (2000). | Article | PubMed | ISI | ChemPort |
-
Parker, P.J. & Parkinson, S.J. AGC protein kinase phosphorylation and protein kinase C. Biochem. Soc. Trans. 29, 860–863 (2001). | Article | PubMed | ISI | ChemPort |
-
Alessi, D.R. et al. Characterization of a 3-phosphoinositide-dependent protein kinase which phosphorylates and activates protein kinase B. Curr. Biol. 7, 261–269 (1997). | Article | PubMed | ISI |
|
neo allele in DN3 and DN4 thymocyte subsets of NLC and LckCre+Pdpk1-/- mice. (b) Total thymocyte numbers from 5- to 7-week-old LckCre+Pdpk1-/- (n = 8) and NLC (n = 6) mice. (c) Flow cytometry of CD4 and CD8 subpopulations in thymocytes from LckCre+Pdpk1-/- and NLC mice. Two-dimensional dot plots of Thy-1-gated cells are representative of four independent experiments. Gates show DN, DP and CD4 populations. (d) Expression of TCR
