Stable silencing of the glucocorticoid receptor in myelin-specific T effector cells by retroviral delivery of shRNA: Insight into neuroinflammatory disease


  • Denise Tischner,

    1. Department of Cellular and Molecular Immunology, University of Göttingen Medical School, Göttingen, Germany
    Search for more papers by this author
  • Jens van den Brandt,

    1. Department of Cellular and Molecular Immunology, University of Göttingen Medical School, Göttingen, Germany
    Search for more papers by this author
  • Andreas Weishaupt,

    1. Department of Neurology, Clinical Research Unit for Multiple Sclerosis and Neuroimmunology, University of Würzburg, Würzburg, Germany
    Search for more papers by this author
  • Fred Lühder,

    1. Institute for Multiple Sclerosis Research, University of Göttingen and Gemeinnützige Hertie-Stiftung, Göttingen, Germany
    Search for more papers by this author
  • Marco J. Herold,

    1. Institute for Virology and Immunobiology, University of Würzburg, Würzburg, Germany
    Current affiliation:
    1. Walter and Eliza Hall Institute of Medial Research, Melbourne
    Search for more papers by this author
  • Holger M. Reichardt

    Corresponding author
    1. Department of Cellular and Molecular Immunology, University of Göttingen Medical School, Göttingen, Germany
    2. Institute for Virology and Immunobiology, University of Würzburg, Würzburg, Germany
    • Cellular and Molecular Immunology, University of Göttingen Medical School, Humboldtallee 34, 37073 Göttingen, Germany Fax: +49-551-395843
    Search for more papers by this author


Autoimmune responses in the CNS can be induced by adoptive transfer of CD4+ T effector cells after antigen-restimulation and expansion of clonal cell lines in vitro. However, pathogenic factors remain partially elusive due to the lack of appropriate methods to achieve gene inactivation. Here we describe a protocol for stable gene silencing in differentiated rat T cells by retroviral transfer of small hairpin RNAs. Through the combination of an expression cassette containing the green fluorescent protein with a puromycin selection cassette this allows for the generation of pure knockdown cell lines suitable for tracking in animals. Exemplified for the glucocorticoid receptor, we demonstrate that gene silencing renders T effector cells unresponsive to ligand-induced apoptosis and gene regulation without affecting their ability to induce EAE in rats. Interestingly, glucocorticoid administration remains effective in the treatment of EAE despite strongly diminished glucocorticoid receptor expression in antigen-specific T cells. This highlights an important role of other cell types and bystander T cells as targets of glucocorticoid therapy. Collectively, our approach provides a simple tool for stable and efficient gene silencing in T effector cells, which should help to better understand brain autoimmune pathophysiology.


MS is a chronic inflammatory disorder of the CNS 1. In the early phase of the disease autoreactive T cells cross the blood–brain barrier (BBB) and become restimulated by local APC. This results in the secretion of cytokines such as IL-17 and IFN-γ, which initiate an inflammatory response by activating microglia and recruiting macrophages. Consequently, the BBB is disrupted allowing for further leukocyte infiltration and deposition of humoral components. Eventually, this leads to oligodendrocyte damage, demyelination and axonal loss 2.

EAE is a frequently used rodent model of MS 3. In Lewis rats, adoptive transfer (AT) of myelin basic protein (MBP)-specific CD4+ T effector cells into syngeneic animals results in a monophasic disease course predominantly reflecting the inflammatory features of MS. So-called encephalitogenic T cells (Tenc) used for EAE induction are initially obtained from regional lymph nodes of antigen-primed animals and subsequently expanded as clonal cell lines by repeated restimulation in vitro4. This offers the opportunity to manipulate the differentiated T effector cells in culture prior to disease induction, e.g. by stable gene delivery of fluorescent marker proteins. Such an approach was successfully applied to the delivery of an expression cassette containing the enhanced green fluorescent protein (eGFP) into antigen-specific T cells and provided new insight into the pathophysiology of brain autoimmune diseases 5, 6. Nevertheless, pathogenic factors essential for the function of fully differentiated autoreactive T effector cells remain partially elusive owing to the lack of appropriate methods for gene inactivation.

The knockout technology is a powerful tool to generate loss-of-function mutations 7. However, it is time-consuming, costly and limited to mice. RNA interference (RNAi) holds great promise as an alternative experimental tool to silence gene expression in almost all cell types and in a variety of species including rats. This method is based on the introduction of so-called siRNA into cells. After binding to the target transcript, the resulting dsRNA complex is incorporated into a large enzyme leading to mRNA degradation 8, 9. Experimentally, stable introduction of siRNA into cells can be achieved by expressing a small hairpin RNA (shRNA) using retroviral or lentiviral vectors. By this means, the state of silencing is inherited over many cell divisions or animal generations 10, 11 and allows for detailed analyses of gene function in vivo.

Glucocorticoids (GC) are in widespread clinical use for the treatment of inflammatory disorders. Acute relapses and optic neuritis in MS patients, for example, are treated by intravenous administration of high GC doses 12, which leads to lymphocyte apoptosis 13, 14, suppression of pro-inflammatory cytokines 15, 16 and an expansion of regulatory T cells 17. Furthermore, GC interfere with leukocyte migration through downregulating adhesion molecules 18, 19 and restoring the integrity of the BBB 20, 21. Recently, we could show that GC actions in T cells but not myeloid cells are essential for therapeutic efficiency in a chronic EAE model in mice 22. However, by use of the employed mouse models we were unable to address the role of the GC receptor (GR) in T effector cells. Therefore, we applied the AT-EAE model in the current approach since this endowed us with the ability to silence GR expression selectively in antigen-specific T effector cells. Our results now show that EAE can be efficiently treated with GC even in the virtual absence of the GR from pathogenic T effector cells. This reveals an important role of bystander T cells and other cell types in the pathogenesis of EAE and GC therapy.


Stable silencing of the GR in T effector cells by RNAi

Antigen-specific T effector cells can be manipulated by retroviral gene delivery without compromising their ability to induce EAE 5. Here we have adopted this approach to selectively silence gene expression in Tenc cells by RNAi. Two different shRNA sequences specific for the rat GR were cloned into the background of the eukaryotic microRNA-30 and expressed under the control of the LTR promoter of the retroviral vector MSCV-LMP. Incorporated as a second component was eGFP and a puromycin selection cassette to allow for the elimination of non-transduced cells (Fig. 1A). Freshly thawed Tenc cells were infected with a highly concentrated preparation of retroviral particles generated with either of the two constructs (siGR-A and siGR-B) or of the empty vector as a control (LMP) resulting in a transduction efficiency of around 50% based on flow cytometric analysis of eGFP expression (Fig. 1B). Following one round of antigen-restimulation, the resulting Tenc cells were cultured in the presence of puromycin to eliminate non-transduced cells. Using this method, we obtained pure cell lines for all three constructs (Tenc LMP, siGR-A and siGR-B) as indicated by the almost exclusive presence of eGFP-positive cells (Fig. 1B and data not shown).

Figure 1.

GR silencing in T effector cells by stable transduction with shRNA-encoding retroviruses. (A) Schematic representation of the retroviral vector MSCV-LMP used for transduction of Tenc cells. The vector contains one cassette consisting of an shRNA embedded in the microRNA-30 background and expressed under the control of the LTR promoter, and a second one consisting of the puromycin resistance cassette (Puro) linked to eGFP by an IRES sequence under the control of the phosphoglycerate kinase promoter. (B) The generation of a stably transduced cell line is exemplified for the empty MSCV-LMP control vector. Tenc cells were infected with retroviral particles, selected in the presence of puromycin and analyzed by flow cytometry for eGFP expression. Histogram plots of uninfected Tenc cells (control), retrovirally transduced Tenc cells (unselected) and Tenc cells obtained after culture in the presence of puromycin (selection) are depicted. (C) Tenc cells stably transduced with the control vector (LMP) and two lines (siGR-A and siGR-B) expressing different shRNA against the rat GR were analyzed for GR protein levels by Western blot; β-tubulin served as a loading control.

GR silencing in T effector cells abolishes ligand-induced transactivation and transrepression

To test whether GR expression was indeed silenced in Tenc siGR-A and siGR-B cells, we analyzed protein levels by Western blot. Expression was reduced by 81 and 71%, respectively, as compared with Tenc LMP cells transduced with the empty vector (Fig. 1C). Thus, retroviral transduction allows for the generation of pure antigen-specific T effector cell lines with diminished GR expression.

To further confirm GR silencing, we performed a number of functional assays. GC-induced apoptosis and upregulation of GILZ (glucocorticoid-induced leucine zipper) both depend on GR-driven transactivation 23, 24. To test this property, we first cultured the Tenc cells for 24 h in the presence of dexamethasone (Dex) and analyzed them for survival by flow cytometry. Although LMP control cells readily underwent cell death, both Tenc siGR cell lines were refractory to GC-induced apoptosis (Fig. 2A). Second, we studied gene expression after 6 h of Dex treatment by quantitative PCR. As predicted, induction of GILZ was reduced although not completely abolished in both Tenc siGR cell lines (Fig. 2B). Taken together, our results indicate that GR-dependent transactivation is strongly impaired.

Figure 2.

Functional analysis of stably transduced Tenc cell lines. (A) Tenc LMP (control cells), Tenc siGR-A and Tenc siGR-B cells (shRNA-expressing cell lines) were cultured in the presence of Dex for 24 h. The relative survival as compared with untreated control cells was determined by AnnexinV/7-AAD staining, n=4–5. Statistics: two-way ANOVA; ***p<0.001. (B–D) GR-dependent gene transactivation and transrepression were analyzed by quantitative RT-PCR. The Tenc cell lines LMP, siGR-A and siGR-B were cultured in the absence (medium control) or presence of 10−6 M Dex for 5 h. The relative induction of GILZ is depicted (B) and the relative repression of IFN-γ (C) and IL-17 (D) compared with their respective medium controls. The expression of each gene was normalized to β-actin, n=3–7. Statistics: Mann–Whitney U-test; *p<0.05 and **p<0.01.

Inhibition of pro-inflammatory cytokines depends on GR-mediated transrepression, an activity that is central to the therapeutic activity of GC 25. IFN-γ and IL-17 are two key cytokines of Tenc cells that are responsible for their effector functions and subject to regulation by Dex. Importantly, GC completely failed to inhibit IFN-γ and IL-17 expression in Tenc siGR-A and siGR-B cells, whereas efficient repression was observed in Tenc LMP control cells (Fig. 2C and D). Thus, GR silencing in Tenc cells renders them largely refractory to GC actions mediated both by transactivation and by transrepression.

GC ameliorate EAE irrespective of GR silencing in T effector cells

Initially, we investigated whether retroviral transduction per se or silencing of the GR impaired the pathogenicity of Tenc cells. Untransduced cells, Tenc LMP, siGR-A or siGR-B cells were adoptively transferred into Lewis rats and the clinical signs monitored over a 10-day period. No differences in the onset, severity or duration of the disease were observed between the four experimental groups (Fig. 3A and B and data not shown). This confirms that T-cell engineering has no impact on the ability of Tenc cells to induce EAE and that the level of GR expression in pathogenic T effector cells does not influence the development or the clinical course of the disease.

Figure 3.

GC therapy of EAE following disease induction by AT of retrovirally transduced Tenc cell lines. (A and B) AT-EAE in Lewis rats was induced by intravenous injection of Tenc cells retrovirally transduced with the empty retroviral vector (Tenc LMP) or vectors expressing a GR-specific shRNA (Tenc siGR-A or Tenc siGR-B). On day 3 after disease induction the animals were intravenously administered PBS (control) or 2 mg/kg Dex (indicated by the arrow), n=5 per group. The experiment was repeated twice with similar results. The clinical disease course is shown over a 10-day period and was found to be significantly different (***p<0.001) for all three cell lines when comparing animals treated with Dex versus PBS between day 3 and 10 (Statistics: two-way ANOVA). (C) On day 4 after EAE induction (disease maximum), infiltrating leukocytes of animals that had received either Tenc LMP, Tenc siGR-A or Tenc siGR-B cells were isolated from the spinal cord and cultured in the presence of Dex for 24 h. The relative survival as compared with the untreated control culture was determined by flow cytometry using AnnexinV/7-AAD staining and gating on eGFP+CD4+ T cells, n=4 per group. Statistics: two-way ANOVA; ***p<0.001.

Next, we asked whether GR silencing in Tenc cells abrogated the therapeutic effect of GC administration on EAE. Rats were intravenously treated with 2 mg/kg Dex on day 3 after disease induction, i.e. when the animals displayed first clinical symptoms. Surprisingly, Dex ameliorated the disease course irrespective of GR silencing in pathogenic T effector cells, i.e. the clinical scores were indistinguishable between rats having received Tenc LMP, siGR-A and siGR-B cells (Fig. 3A and B). Thus, functional inactivation of the GR in antigen-specific T effector cells does not compromise the therapeutic efficacy of GC therapy.

To exclude the possibility that the retrovirally introduced shRNA was no longer functional after AT, we isolated lymphocytes at the disease maximum (day 4 after transfer) from the spinal cord of animals having received Tenc LMP or Tenc siGR cells. Subsequently, the cells were assessed ex vivo for their ability to undergo apoptosis in response to Dex. Importantly, CNS-infiltrating Tenc siGR cells were still refractory to GC treatment, confirming that the transgene remains active in vivo (Fig. 3C).

GR-deficient T effector cells in peripheral lymphoid organs are resistant to GC

We had previously reported that Dex therapy of EAE primarily induces apoptosis in peripheral T cells. To test this feature we injected Tenc LMP or siGR-B cells into Lewis rats and treated them with 2 mg/kg Dex on day 3. Another 20 h later the animals were sacrificed and the T-cell subsets enumerated in spleen. GC administration strongly reduced the number of antigen-unspecific T cells (eGFPCD4+ and eGFPCD8+ T cells) in both experimental groups as compared with PBS treated animals (Fig. 4A). In contrast, Dex treatment only diminished the number of Tenc LMP but not Tenc siGR-B cells (Fig. 4B). This strongly suggests that silencing of the GR renders pathogenic T effector cells in peripheral lymphoid organs resistant toward GC-induced apoptosis in vivo.

Figure 4.

Enumeration of antigen-unspecific and MBP-specific Tenc cells in the spleen following Dex therapy. AT-EAE in Lewis rats was induced by intravenous injection of Tenc LMP or Tenc siGR-B cells. On day 3 after disease induction animals were treated intravenously with PBS (control) or 2 mg/kg Dex and 20 h later T-cell numbers in spleen were determined by flow cytometry. The number of antigen-unspecific eGFP- T cells (A) and eGFP+ Tenc effector cells (B) in spleen after Dex therapy are depicted relative to PBS treated control rats that received the same cell line, n=5 per group. The experiment was repeated twice with similar results. Statistics: Mann–Whitney U-test; *p<0.05 and n.s., not significant.

GC improve pathophysiological features in the CNS despite GR silencing in T effector cells

Our earlier studies indicated that GC effects on leukocytes in CNS lesions are mainly indirect and reflect events taking place in peripheral lymphoid organs 22. Therefore, we studied the inflammatory infiltrate in the spinal cord after Dex treatment. In rats having received Tenc LMP control cells the numbers of infiltrating eGFP+ Tenc cells (Fig. 5A), eGFP bystander T cells (Fig. 5B) and macrophages (Fig. 5C) were strongly reduced after GC administration. In addition, the integrity of the BBB was partially restored as indicated by a weaker albumin staining of the spinal cord (Fig. 5D). This correlates with the observed improvement of the clinical disease course in this experimental group (see Fig. 3). However, when we used Tenc siGR-B cells to induce EAE, the number of infiltrating T effector cells was only marginally decreased by GC therapy (Fig. 5A). This is consistent with the strongly reduced level of GR expression in Tenc siGR-B cells (see Fig. 1C). In contrast, Dex treatment efficiently reduced the number of bystander T cells and macrophages in these animals (Fig. 5B and C). Similarly, the integrity of the BBB was almost equally well restored by GC therapy in rats of both experimental groups (Fig. 5D). Nevertheless, the GC effects on the non-silenced cell-types were slightly less pronounced as compared with those observed in rats having received Tenc LMP cells, indicating that GR inactivation in T effector cells may indirectly also impact on other cell types involved in the pathogenesis of EAE.

Figure 5.

Analysis of the spinal cord infiltrate and the BBB integrity after GC therapy. AT-EAE was induced in Lewis rats by injecting Tenc LMP or Tenc siGR-B cells. On day 3 after disease induction animals were treated intravenously with PBS (control) or 2 mg/kg Dex and 20 h later a piece of the cervical spinal cord was analyzed by histology, whereas infiltrating leukocytes were purified from the rest by density centrifugation. (A and B) The abundance of eGFP+ Tenc effector cells (A) and eGFP bystander T cells (B) was determined by flow cytometry. (C) The number of infiltrating macrophages was determined by counting serial immunohistochemical stainings and their relative numbers in Dex-treated animals depicted as percentages in relation to the respective PBS-treated rats that received the same cell line. (D) The integrity of the BBB was analyzed by albumin staining, subsequently quantified by densitometry and depicted in arbitrary units in relation to the respective PBS treated rats that received the same cell line. A reduction in optical density corresponds to restored BBB integrity, n=5 per group. The experiment was repeated twice with similar results. Statistics: Mann–Whitney U-test; *p<0.05 and n.s., not significant.

It is noteworthy that we did not observe any difference in the cellular composition of the CNS infiltrate irrespective of whether EAE was induced by AT of Tenc LMP or retrovirally unmanipulated cells (data not shown). This indicates that T-cell engineering does not influence leukocyte infiltration and confirms that the observed effects were indeed the consequence of GR silencing. Taken together, we conclude that modulation of pathogenic T effector cells is not essential for GC therapy of EAE and presumably MS.


The use of knockout mice allowed multiple new insights into the pathophysiology and therapy of brain autoimmune diseases 22, 26, 27. Notwithstanding the great benefits of this technology, it also has some drawbacks. In this respect, RNAi in combination with retroviral vectors holds great promise as an alternative experimental tool. First, gene inactivation in knockout mice usually occurs at an early stage, which precludes distinguishing effects on development, differentiation and effector functions of cells. In contrast, stable gene silencing by RNAi can also be achieved in fully differentiated cells by using retroviral vectors. Second, the knockout technology is limited to mice since homologous recombination in ES cells is not yet available in other species. Therefore, the advantage of certain rat models of brain autoimmune diseases 28 can only be exploited by generating knockdown animals 10, 11 or by AT of cells after RNAi-mediated gene silencing. Third, the generation of knockout mice is costly and time-consuming, for which reason it does not qualify for large-scale screenings. In contrast, the approach taken here offers the unique possibility to silence several genes in parallel followed by functional analysis. Finally, knockout mice are inconclusive when gene duplication has occurred, e.g. inactivation of one subtype of the anti-apoptotic protein A1 resulted in a very subtle phenotype 29. This problem can be overcome by concomitantly silencing several isoforms of a gene using a single siRNA-targeting homologous region 30. Nevertheless, despite its favorable characteristics RNAi has also a number of disadvantages. Above all, gene silencing always remains incomplete whatsoever the efficiency of inactivation will be. Thus, effects of residual gene expression can never be excluded. In addition, it has been reported that shRNA may cause off-target effects and could lead to a saturation of the RNAi machinery, resulting in an impaired processing of endogenous microRNA 31, 32. Taken together, stable gene silencing by retroviral transfer of siRNA is a valuable tool that bears the great potential to complement the knockout technology in answering selected scientific questions.

A major obstacle in the application of RNAi to immunological questions is the fact that rodent lymphocytes are particularly resistant to siRNA-mediated downregulation of gene expression. Primary mouse T cells are poorly susceptible to retroviral, lentiviral as well as adenoviral delivery of shRNA constructs 33. In contrast, transfection efficiency of naïve CD4+ T cells could be increased to around 50% by using the nucleofection method 33. Importantly, the activation and developmental status of lymphocytes also strongly impacts the success of gene inactivation by RNAi. In support of this notion, CD4 expression could be silenced in proliferating T cells from D011.10 mice stimulated with OVA peptide while naïve lymphocytes were refractory 34. Furthermore, studies by Rajewsky and collegues revealed that in contrast to developing thymocytes gene silencing in mature lymphocytes was particularly poor 30. In the light of these observations, the method described here represents an interesting alternative. By combining a highly efficient retroviral transduction protocol with the ability to select transformed cells in the presence of antibiotics, CD4+ T lymphocytes can be obtained in which almost all cells harbor the shRNA-expressing construct. In addition, this technique results in stable gene silencing, which is maintained over several weeks in cell culture and even after transfer into animals in the context of an inflammatory disease. To our knowledge, those features are not achieved by other methods presently available, for which reason it may prove highly valuable to address selected scientific questions.

Not withstanding the challenges encountered when applying RNAi to immunological questions, this technique had been previously applied to inactivate cytokines, their receptors, down-stream signaling molecules and transcription factors that direct the expression of inflammatory genes. Silencing of TNF-α, for example, was achieved by electroporation of a siRNA directly into the joint, resulting in improved disease symptoms in a model of rheumatoid arthritis 35. Using a different method, expression of the GM-CSF receptor was inhibited in hematopoietic stem cells by lentiviral transduction 36. Furthermore, the development of EAE was inhibited by silencing T-bet, a transcription factor that drives T-cell differentiation into a TH1 phenotype, through transfection of antigen-specific T cells with synthetic oligonucleotides 37. However, none of those approaches had allowed stable and inheritable gene silencing similar to the strategy described here for antigen-specific T effector cells. Such clonal cell lines can be isolated from Lewis rats immunized with the myelin-antigen MBP, propagated in vitro over weeks and used to induce EAE. Here we generated MBP-specific T effector cell lines in which expression of the GR was selectively silenced. In contrast to knockout mice, this allowed inactivating the GR in T lymphocytes after differentiation into effector cells had already occurred. Taking advantage of a puromycin-resistance cassette in the vector and the eGFP marker protein, we obtained almost pure knockdown cell lines with stably reduced gene expression. Inactivation of the GR was confirmed for two cell lines generated with different shRNA sequences by Western blot analysis and functional studies. Protein levels of the GR were diminished by around 80%, transactivation of the GC response element-dependent gene GILZ was reduced, GC-induced apoptosis was abolished and repression of pro-inflammatory cytokines by Dex was completely abrogated. Moreover, the ability of the transduced cells to induce EAE after AT was indistinguishable from control cells and the shRNA remained functionally active in vivo. Thus, stable expression of siRNA allows for efficient gene inactivation in Tenc cells without compromising their pathogenicity.

GC are widely used to treat acute relapses in MS patients by exploiting their potent pro-apoptotic and anti-inflammatory activity 38, 39. Previously we had identified peripheral T lymphocytes as essential targets but the specific role of GC actions on T effector cells remained elusive. T-cell-specific knockout mice as well as mice lacking the GR in all hematopoietic cells were refractory to Dex treatment in the chronic MOG-EAE model, whereas disruption of the GR in myeloid cells did not impact therapeutic success 22. Furthermore, Dex induced apoptosis in splenic T cells, impaired their migratory capacity but did not exert any direct effect on T lymphocytes in the CNS lesion 22. Keeping with these results, Dex administration in the rat AT-EAE model reduced T-cell numbers in spleen although it had no influence on T effector cells lacking GR expression. Furthermore, GC treatment diminished the number of T effector cells, bystander T cells and macrophages in the spinal cord and partially restored the integrity of the BBB. Qualitatively similar effects with regard to CNS infiltration and BBB disruption were seen when Tenc LMP or Tenc siGR cells were used for disease induction, with the exception that GR-deficient T effector cells in the infiltrate were again largely refractory to GC therapy. Importantly, amelioration of EAE after Dex treatment correlated with the improved pathophysiological features observed in both experimental groups. This provides evidence that repression of Tenc cell function is not a prerequisite for successful GC therapy of EAE.

At first sight, successful Dex treatment of EAE after GR silencing in T effector cells might seem contradictory to our recent report that T cells are essential targets for GC therapy of EAE 22. However, one has to keep in mind that gene inactivation in T-cell-specific GR knockout mice abolishes the effects of Dex treatment in both, antigen-specific T effector as well as bystander T cells attracted to the inflamed CNS 22. In contrast, the strategy used here limits silencing of the GR to pathogenic T lymphocytes with a well-defined antigen-specificity, whereas bystander T cells remain fully responsive to GCs. Thus, the different outcomes obtained in these two experimental systems suggest that modulation of non-encephalitogenic T cells significantly contribute to the therapeutic efficiency of GC. Nevertheless, differences between the chronic MOG-EAE model induced by active immunization of C57Bl/6 mice and the monophasic AT-EAE in Lewis rats cannot be fully ruled out as additional influential factors. Therefore, it will be interesting in the future to study T effector cell-specific GR inactivation also in other model systems.

In summary, we have developed a simple tool for efficient long-term gene silencing in T effector cells, which we expect to become instrumental in elucidating factors that are required for T lymphocyte function in brain autoimmune diseases following their differentiation into effector cells.

Materials and methods


Lewis rats were purchased from Charles River (Sulzfeld, Germany) and used at 6–10 wk of age. All experiments were conducted according to Lower Saxonia state regulations for animal experimentation and approved by the responsible authorities (Az: 33.42502-04-020/07, Nds. Landesamt für Verbraucherschutz und Lebensmittelsicherheit).

T-cell culture

Antigen-specific T effector cell lines (Tenc cells) were established by the immunization of Lewis rats with guinea pig (gp) MBP followed by antigen restimulation in vitro as described 4, 40. The cells were cultured in RPMI-1640 medium supplemented with 1% rat serum, 2 mM glutamine and standard antibiotics. Prior to disease induction, 3×106 gpMBP-specific T-cell blasts were freshly thawed and activated for 3 days using 1.5×108 irradiated (30 Gray) thymocytes and 10 μg/mL gpMBP 41. At the end of the culture period, the Tenc cells were purified by gradient centrifugation, resuspended in PBS and used for injection. To study Tenc cell function in vitro or to analyze lymphocytes isolated from the spinal cord, we cultured them in RPMI-1640 medium supplemented with 10% FCS and standard antibiotics for up to 24 h in the presence or absence of Dex.

Cloning of retroviral vectors

The two rat GR-specific shRNA sequences cloned in the retroviral vector pSM2 were obtained from Open Biosystems (Huntsville, AL, USA). siRNA-A: tgctgttgacagtgagcga CAGCTTTACATGCAATTTATTTA gtgaagccacagatg TAAATAAATTGCATGTAAAGCTG ctgcctactgcctcgga; siRNA-B: tgctgttgacagtgagcgc GCTCCTGATCTGATTATTAATTA gtgaagccacagatg TAATTAATAATCAGATCAGGAGC atgcctactgcctcgga. The shRNA were excised from the pSM2 vector and inserted into the microRNA-30 background of the retroviral vector MSCV-LMP via its EcoRI and XhoI restriction sites (also Open Biosystems). A database search revealed that the maximal complementarity of the two selected siRNA sequences with other rat mRNAs was 14 (siRNA-A) and 15 nt (siRNA-B), respectively, (

Production of retroviral particles and transduction of T effector cells

For the generation of retroviral particles 1.3×106 ecotropic Phoenix producer cells were seeded in 6 cm dishes and cultured in DMEM supplemented with 10% FCS and 1% standard antibiotics. On the following day, the medium was changed to HEPES-buffered DMEM and 4 h later the cells were transfected with 10 μg retroviral vector DNA by CaPO4 precipitation following standard protocols. For each individual retroviral construct 60 plates were transfected in total. One day later, the medium was changed and after 48 h the retrovirus-containing supernatant was harvested, combined and filtrated. Afterwards, the virus preparation was concentrated by ultracentrifugation (4°C, 5500×g, 20 h) and resuspended in a final volume of 4 mL DMEM medium.

In total 3×106 freshly thawed gpMBP-specific T effector cells were transduced in the presence of 10 μg/mL polybrene with the complete 4 mL retroviral concentrate by spinoculation (32°C, 870×g, 3 h). Subsequently, the cells were resuspended in fresh cytokine-containing medium, cultured for 3 days and then restimulated with irradiated thymocytes and gpMBP (see section T-cell culture). Selection for stable provirus integration was achieved by culturing the transduced T effector cells in the presence of 2.5 μg/mL puromycin for 3 days. Purity was assessed by flow cytometry and usually greater than 98% (see also Fig. 1B). Finally, the new cell lines were expanded for several rounds by repeated antigen restimulation to obtain sufficiently large numbers for EAE induction and functional analyses. Viability and continuous transgene expression were always analyzed before use.

Isolation of spinal cord infiltrates

Lymphocytes were isolated from the spinal cord by density centrifugation following perfusion with PBS. The dissected tissue was passed through a metal mash and homogenized in PBS containing 0.1% BSA, 1% glucose, and 100 μg/mL DNase I (Roche Diagnostics, Mannheim, Germany). After centrifugation, the spinal cord homogenate was resuspended in 6 mL of 30% Percoll, overlaid on a Percoll gradient containing 4 mL of 45% and 2 mL of 70% Percoll and spun for 20 min (2300 rpm, 4°C). Finally, the lymphocytes were harvested at the interfaces between the layers, washed with PBS and analyzed by flow cytometry.

Induction of AT-EAE

AT-EAE was induced by injecting 7×106 gpMBP-specific CD4+ T effector cells into the tail vein of Lewis rats (Tenc LMP, Tenc siGR-A or Tenc siGR-B). Animals were weighted daily and inspected for signs of EAE. The severity of EAE was assessed employing a six-grade disease score as follows: 0 = healthy; 1 = weight loss, limp tip of tail; 2 = limp tail, mild paresis; 3 = moderate paraparesis, ataxia; 4 = tetraparesis; 5 = moribund and 6 = dead 42, 43.

GC therapy was performed by injecting 2 mg/kg Dex (Sigma-Aldrich, Taufkirchen, Germany) or PBS as a control into the tail vein on day 3 after disease induction, i.e. when animals displayed first signs of disease as indicated by a clinical score of 1–2. Twenty hours later the animals were sacrificed and analyzed by flow cytometry and histology. To follow the clinical disease course and the change in body weight over time, some animals from each group were left alive during the whole experiment until they had fully recovered from the disease.

Flow cytometry

All Ab and reagents used for FACS analyses were obtained from BD Biosciences unless otherwise indicated: Ox35 (CD4), Ox38 (CD4), Ox39 (CD25), Ox40 (CD134), R73 (βTCR), WT.1 (CD11α), AnnexinV, 7-AAD. The Abs or reagents were directly labeled with FITC, PE, PerCP, PE-Cy7, Cy5, APC or APC-Cy7. Extracellular staining was performed as described previously 44. All analyses were performed on a FACSCanto II device allowing for the detection of six fluorescent dyes (BD Biosciences, Heidelberg, Germany).

Histological analysis and immunohistochemistry

Samples from the cervical spinal cord were embedded in paraffin, 5 mm sections mounted on poly-L-lysine-coated slides and processed as described 45. Pretreatment with hydroxylamine (0.9%, Sigma-Aldrich) was required for the albumin staining. Immunohistochemistry was performed as detailed previously 45. In brief, serial sections were stained with the monoclonal Ab ED-1 (Serotec, Düsseldorf, Germany) for the detection of macrophages or with an anti-albumin Ab (Nordic, Tilburg, Netherlands) to analyze the integrity of the BBB. Spinal cord sections were examined in a blinded manner to determine the number of macrophages; albumin stainings were quantified by gray-scale analysis using the MetaVue program (version 5.3r2, Visitron Systems, Munich, Germany) and the resulting pixel densities depicted as arbitrary units.

Quantitative PCR and Western blot analysis

Total RNA was isolated using TriZol reagent (Invitrogen, Karlsruhe, Germany), treated with DNaseI (Roche Diagnostics) and purified by use of the RNaesy mini Kit (Qiagen, Hilden, Germany). In total 500 ng of RNA were transcribed into cDNA with Superscript II Reverse Transcriptase (Invitrogen). Real-time PCR was performed on an ABI7500 instrument (Applied Biosystems) using a SYBRgreen mastermix followed by normalization to β-actin expression using the relative quantification method. Primer sequences are available upon request. PCR conditions were as follows: 95°C denaturation for 2 min followed by 40 cycles of 95° for 15 s and 60° for 45 s.

Protein extracts were prepared in radioimmune precipitation assay lysis buffer, separated on a 10% SDS-PAGE gel, transferred to a polyvinylidine difluoride membrane, and detected using the following Ab: anti-GR (M-20; Santa Cruz, Heidelberg, Germany), anti-β-tubulin (cloneTUB2.1, Sigma-Aldrich). Quantification of band intensities was achieved using the GelPro Analyzer 4.5 software (Media Cybernetics).

Statistical analysis

All data are depicted as mean±SEM. Analysis was performed by using the Mann–Whitney rank sum test or a two-way ANOVA as indicated in the text (Statistica 6.0, Statsoft GmbH). *p<0.05, **p<0.01 and ***p<0.001 were considered as significant p-values.


The authors thank Amina Bassibas and Sabrina Braunschweig for excellent technical help, Chi Wan Ip for densitometric analysis of histological sections and Dr. Thomas Hünig for critical reading of the manuscript. This work was supported by grants from the Gemeinnützige Hertie-Stiftung (1.01.1/06/010, to F.L. and H.M.R.) and the DFG (Re1631/1-3, to H.M.R.).

Conflict of interest: The authors declare no financial and commercial conflicts of interest.