|Home | About | Journals | Submit | Contact Us | Français|
Ocular infection with herpes simplex virus 1 (HSV-1) sets off an inflammatory reaction in the cornea which leads to both virus clearance and chronic lesions that are orchestrated by CD4 T cells. Approaches that enhance the function of regulatory T cells (Treg) and dampen effector T cells can be effective to limit stromal keratitis (SK) lesion severity. In this report, we explore the novel approach of inhibiting DNA methyltransferase activity using 5-azacytidine (Aza; a cytosine analog) to limit HSV-1-induced ocular lesions. We show that therapy begun after infection when virus was no longer actively replicating resulted in a pronounced reduction in lesion severity, with markedly diminished numbers of T cells and nonlymphoid inflammatory cells, along with reduced cytokine mediators. The remaining inflammatory reactions had a change in the ratio of CD4 Foxp3+ Treg to effector Th1 CD4 T cells in ocular lesions and lymphoid tissues, with Treg becoming predominant over the effectors. In addition, compared to those from control mice, Treg from Aza-treated mice showed more suppressor activity in vitro and expressed higher levels of activation molecules. Additionally, cells induced in vitro in the presence of Aza showed epigenetic differences in the Treg-specific demethylated region (TSDR) of Foxp3 and were more stable when exposed to inflammatory cytokines. Our results show that therapy with Aza is an effective means of controlling a virus-induced inflammatory reaction and may act mainly by the effects on Treg.
IMPORTANCE HSV-1 infection has been shown to initiate an inflammatory reaction in the cornea that leads to tissue damage and loss of vision. The inflammatory reaction is orchestrated by gamma interferon (IFN-γ)-secreting Th1 cells, and regulatory T cells play a protective role. Hence, novel therapeutics that can rebalance the ratio of regulatory T cells to effectors are a relevant issue. This study opens up a new avenue in treating HSV-induced SK lesions by increasing the stability and function of regulatory T cells using the DNA methyltransferase inhibitor 5-azacytidine (Aza). Aza increased the function of regulatory T cells, leading to enhanced suppressive activity and diminished lesions. Hence, therapy with Aza, which acts mainly by its effects on Treg, can be an effective means to control virus-induced inflammatory lesions.
Once a viral infection becomes established, its removal largely depends on the activity of T lymphocytes. Multiple functional subsets of T cells can participate with the outcome, dependent on the nature of the virus, its location in the body, and the types of T cells that become activated and expanded by the infection (1, 2). Chronic tissue-damaging inflammatory reactions can occur when elimination of infection is difficult to achieve or the balance of T cell responsiveness emphasizes proinflammatory cells that contribute to tissue damage (3). For example, in stromal keratitis (SK) resulting from ocular infection by herpes simplex virus 1 (HSV-1), a chronic inflammatory reaction occurs in the corneal stroma that is orchestrated mainly by proinflammatory CD4 Th1 and Th17 T cells (4,–6). The lesion is less severe and can even resolve if regulatory T cells (Treg), such as Foxp3+ CD4 T cells, are dominant over the other proinflammatory CD4 T cell subsets (7,–11). Accordingly, therapies aimed at increasing Treg numbers and/or improving their regulatory activity are of high relevance.
It is becoming evident that the balance between inflammatory and regulatory T cells is not fixed but can change as a consequence of one or the other cell type changing in number or altering its functional activity (12). For example, functional changes were observed in vitro when Treg were exposed to some inflammatory mediators (13, 14). Similar functional changes may occur during autoinflammatory lesions in vivo, with Treg losing their regulatory activity (15, 16). Of more concern, these Treg may change and take on a proinflammatory function and then contribute to the severity of tissue damage (15,–18). The changes in functional phenotype that occur are likely explained by epigenetic changes that affect the expression of the Treg transcription factor Foxp3 (19, 20). These epigenetic changes usually occur in the highly conserved intron 2, also known as Treg-specific demethylation region (TSDR), which harbors cytosine-phospho-guanine (CpG) sites that are subject to methylation (21, 22). Thus, when TSDR is demethylated, the transcription factors Ets-1 and Creb can bind and act as enhancers for the continuous transcription of the Foxp3 gene (23, 24). However, when the TSDR is methylated, the enhancer activity is diminished and only transient expression of Foxp3 occurs. Consequently, for the stable expression of the Foxp3 gene, demethylation of TSDR is required (25, 26). In fact, natural Treg that derive from the thymus are stable and their TSDR is invariably demethylated (22). In contrast, in vitro- or in vivo-induced Treg have a TSDR which is methylated, and such cells show plasticity of both phenotype and function (22, 25, 27, 28).
Approaches to promote the stability of induced Treg are to block TSDR methylation or to generate Treg that have a demethylated TSDR profile. The latter can be achieved by inhibiting DNA methyltransferase, as occurs when 5-azacytidine (Aza) is used for therapy (22, 25, 29). This FDA-approved drug is used to treat myelodysplastic syndrome (30) and is also an effective therapy against some inflammatory disease models (31,–33). The treatment has also been proposed to act by increasing the potency of the Treg response (32,–34), and we further evaluate this notion using an infectious disease model of inflammation.
In this report, we show that the therapeutic administration of Aza was highly effective at suppressing the severity of ocular immunoinflammatory lesions that result from corneal infection with HSV. The beneficial outcome of the Aza therapy appeared to be the consequence of restricted infiltration of proinflammatory immune cells to the cornea. The cells that did enter had increased numbers of Treg compared to the numbers of CD4+ gamma interferon-producing (IFN-γ+) effector T cells. This increased representation of Treg in Aza-treated animals was also evident in the blood and lymphoid tissues. Significant differences in the suppressive efficacy of Treg from control and treated groups were also observed, with Treg from Aza-treated animals being more suppressive, a property explained at least in part by higher expression levels of reactive oxygen species (ROS) and activation markers. Furthermore, Treg generated in vitro in the presence of Aza expressed a fully demethylated TSDR, and these cells also displayed enhanced suppressive activity, which correlated with the enhanced ROS production and activation markers. Overall, our results emphasize that the epigenetic-modification drug Aza may represent a novel approach to control HSV-1-induced ocular immunopathological lesions, a common cause of infectious blindness in humans in the United States (35).
To assess the efficacy of Aza in reducing the extent of ocular lesions caused by HSV infection, animals were given either Aza or phosphate-buffered saline (PBS; control) daily starting on day 5 postinfection (p.i.). This is the time point when there is at best minimal replicating virus detectable in the infected corneas and early inflammatory reactions start to become evident (36). Animals were examined at intervals to record the severity of SK lesions. The results were clear-cut, with animals receiving Aza therapy showing significantly (P < 0.001) reduced SK lesion severity compared to that in PBS-treated control animals (Fig. 1A) Treatment effects were first evident by day 10, and by day 15, 10% of Aza-treated animals showed a lesion score of ≥3, compared to 60% in PBS-treated control animals (Fig. 1B). This pattern of reduced inflammatory reaction in Aza-treated animals was also evident in histological sections of corneas taken from animals terminated at day 15 p.i. (10 days after treatment) (Fig. 1C).
At the termination of the experiments on day 15 p.i., pools of 4 corneas were collected and processed to identify their cellular composition by fluorescence-activated cell sorting (FACS) analysis. There were reductions in inflammatory cell numbers, including neutrophils (>500-fold), macrophages (10-fold), and CD4 T cells (>10-fold), in Aza-treated animals compared to the numbers in controls (Fig. 2A to toC).C). In separate experiments of the same design, pools of corneas were processed to quantify mRNA of selected cytokines (interleukin 1β [IL-1β], tumor necrosis factor alpha [TNF-α[, IL-6, and IL-12) and chemokines (CCL3, CCL2, CXCL1, and MMP1) by quantitative real-time PCR (qRT-PCR). As shown by the results in Fig. 3, the corneas of mice treated with Aza showed reductions in the levels of several proinflammatory cytokines and chemokines compared to those of controls. However, there were also reductions in the expression levels of anti-inflammatory IL-10 and transforming growth factor β (TGF-β) (Fig. 3), likely explained by reduced numbers of infiltrating immune cells. Taken together, our results show that daily administration of Aza starting 5 days after virus infection significantly diminished HSV-1-induced immunopathology.
Since it is known that the outcome of SK lesion severity is dependent on the ratio of Treg to Th1 (37), the ratios of the cell types were compared in corneas, blood, and draining lymph nodes (DLN) in Aza-treated and control infected animals. Pools of corneas were collected at 15 days p.i. from Aza-treated and control animals, and the infiltrating cell population was recovered after collagen digestion. These cells were then stimulated in vitro for 4 h with phorbol myristate acetate (PMA) and ionomycin, and the CD4 T cells that were either IFN-γ producers or expressed the transcription factor Foxp3 were enumerated (Fig. 2D). Changes in the representation of the two cell types occurred as a consequence of Aza therapy. Thus, in the corneas of Aza-treated animals, the ratio favored Foxp3+ CD4 T cells over Th1 cells, being around 2:1, but in controls, the ratio was around 1:7 (Fig. 2E). Approximately 7% of the total CD4 T cells in the corneas of control animals expressed Foxp3, but in Aza-treated animals, around 35% were Foxp3+ cells (Fig. 2D). As with the corneas, blood from Aza-treated mice displayed an increased Treg frequency, with 12% of the CD4 T cells being Foxp3+, compared to only 5% in PBS-treated controls at day 15 p.i. (Fig. 4A). The total number of CD4 T cells (per 1 million cells) in the blood was decreased by twofold in the Aza-treated animals (Fig. 4B). Although less in magnitude, similar reductions in Th1 frequency and number (Fig. 4D and andE),E), along with changes in the Treg-to-Th1 ratio, occurred in the DLN at day 15 p.i. (Fig. 4F). The numbers of total CD4 T cells and Treg cells in the DLN of the Aza-treated animals were reduced by 1.8- and 1.4-fold, respectively (Fig. 4C and andG).G). The reductions in Treg numbers may be a consequence of reduced inflammatory response. Additionally, single-cell suspensions of DLN isolated at day 15 p.i. from control and Aza-treated animals were stimulated overnight with UV-inactivated HSV-1, followed by intracellular cytokine staining (ICS) assay to measure antigen-specific Th1 responses. The results indicated that there were more than twofold reductions in the frequency and number of Th1 cells (CD4+ IFN-γ+ cells) in DLN samples from Aza-treated animals that were specific for HSV-1 (Fig. 4H). In conclusion, upon Aza treatment, there was a change in the balance in CD4 T cell responses, with an increased representation of Treg, which could in part contribute to the reduced lesion severity.
Since Aza treatment of HSV ocularly infected mice resulted in diminished lesions and increased representation of Foxp3+ CD4 T cells, the question arose as to whether or not the Treg population showed any changes in regulatory function in response to Aza therapy. To measure the suppressive activity, Foxp3-GFP (green fluorescent protein) mice were infected with HSV-1, and Foxp3+ T cells were recovered by FACS from the pools of DLN and spleens of control and Aza-treated animals at day 15 p.i. Equal numbers of Foxp3+ T cells were cultured at different ratios, with naive responders being stimulated with anti-CD3/CD28 antibodies (Abs). The results indicate that Treg from HSV-1-infected, Aza-treated mice displayed more suppressive activity against stimulated naive responders (>20-fold at 1:8 ratio) than did Treg from infected controls (Fig. 5A). Suppression with Treg from Aza-treated animals could be observed in cultures with 1 Treg to 16 responders, whereas with control Treg, a ratio of 1 to 4 was needed to demonstrate significant levels of suppression (Fig. 5B).
To provide an explanation for the greater suppressive activity of cells from Aza-treated infected animals, a number of measurements were made. To determine whether differences in IL-10 production were an explanation, DLN from Aza-treated and control animals were isolated at day 15 p.i., 1 million single-cell suspensions were stimulated with PMA-ionomycin, and supernatants were compared for levels of IL-10 using enzyme-linked immunosorbent assay (ELISA). No significant differences were detectable between the cells from Aza-treated and control groups (data not shown). Comparisons were also made between Treg in control and Aza-treated animals for activation markers, as well as ROS production, since the latter has been proposed to be involved in suppressive activity (38). In the DLN from Aza-treated animals isolated at day 15 p.i., Treg displayed only modestly increased expression (ranging from 1.3- to 1.6-fold) of activation markers. These included CD25, OX40, GITR, CD103, FR4, and CD44 on Foxp3+ CD4 T cells (Fig. 5C). The differences between the two Treg populations were greatest in the case of the expression of intracellular ROS. Thus, the Treg in DLN from Aza-treated animals had around a threefold increase in ROS activity compared to that in the cells from controls (Fig. 5D and andE).E). Additionally, the expression levels of genes involved in ROS production, i.e., NOX-2 and NCF-1 genes (components of NADPH oxidase complex) (39), were also increased 3- and 2.5-fold, respectively, in the Treg of Aza-treated animals compared to those of control animals, as measured by qRT-PCR (Fig. 5F). The expression of IL-10 and TGF-β was also measured in Treg from Aza-treated and control animals using qRT-PCR. The results indicate similar expression levels of IL-10 and TGF-β in both samples (Fig. 5G). Based on the above-described results, we could show that Aza treatment protected the phenotype and function of Treg cells. In conclusion, Treg show enhanced suppressive activity after Aza treatment, and this may be explained at least in part by their increased activation markers and ROS-producing ability.
Since Aza treatment reduced lesion severity, which correlated with changes in Treg number and function, experiments were done to determine the outcome of Aza treatment wherein Treg were depleted prior to infection. Depletion was achieved by the administration of monoclonal antibody (MAb) against the IL-2 receptor (CD25), given on day 0 of infection. The depletion procedure, upon measurement at day 15 p.i., was shown to be around 50% effective at reducing total Foxp3+ T cells in DLN (Fig. 6A). SK lesion severity was measured at day 15 p.i., and the results indicate that Aza treatment of Treg-intact animals led to reduced lesion severity, with an average SK score of 1.7. This compared to an average score of 3.1 in the control groups. However, in animals depleted of Treg and treated with Aza, the inhibitory effect on SK severity was no longer apparent, with the average score being 3.8 (Fig. 6B). To measure any effect of Aza therapy on the magnitude of CD4 Th1 response, the numbers of IFN-γ-producing CD4 T cells in the DLN at day 15 p.i. were measured. Unlike in Treg-intact animals, where Aza treatment resulted in reduced effector T cell numbers, Aza treatment of Treg-depleted animals resulted in no significant difference in effector responses compared to the response in PBS-treated controls. (Fig. 6C).
Next, to evaluate the effect of Aza on the proliferation of Treg and effectors, DLN were isolated at day 15 p.i. from Treg-depleted and control animals treated with or without Aza. Single-cell suspensions were stained for CD4, Foxp3, and Ki-67 (proliferation marker). The results indicated that after Aza treatment, the proliferation of effector T cells was reduced by 1.5-fold in the Treg-intact animals compared to their proliferation in the untreated controls. Whereas Aza treatment in Treg-depleted animals resulted in a 1.5-fold increase in the proliferation of effector T cells compared to that in untreated controls (Fig. 6D), the proliferation of Treg was unchanged after Aza treatment in both cases (Fig. 6E). Consistent with the role of Treg in controlling effector cell proliferation, Aza treatment in the Treg-depleted animals increased the proliferation of effector cells, whereas in the presence of Treg, Aza treatment led to reduced proliferation. Accordingly, our results imply that Aza may act preferentially on the Treg subset that likely expresses a high level of CD25.
To evaluate the direct effects of Aza on Treg and T effectors (Th1), in vitro differentiation experiments were performed. For this purpose, naive splenocytes from DO11.10 RAG2−/− animals (ova peptide-specific and 98% naive CD4+ T cells) were cultured in the presence of Treg-differentiating conditions (IL-2 and TGF-β), as well as in the presence or absence of graded amounts of Aza (from 1 μM to 15 μM). The results show a dose-dependent enhancement in Treg differentiation compared to that in control cells without Aza, with the maximal effect evident at 5 μM (Fig. 7A). This dose yielded an approximately twofold increase in the frequency of Foxp3+ CD4 T cells induced in cultures (Fig. 7B). Similarly, when naive DO11.10 RAG2−/− splenocytes were cultured in the presence of 5 μM Aza and Th1-differentiating conditions (IL-12 and anti-IL-4 Ab), Aza increased the frequency of IFN-γ by twofold (Fig. 7C). To provide a possible explanation for the enhancing effects of Aza on Treg induction, experiments were done to record epigenetic changes in the TSDR region of Treg induced in the presence or absence of Aza. Although Aza might affect the global methylation status of several other genes with CpG sites, such as the GITR, Ctla4, Ikzf4, and CD25 genes (27), the methylation status of only the TSDR region of the Foxp3 gene was evaluated, as this region is known to be an indicator of Treg stability and function (25,–27, 40, 41). Naive CD4 T cells isolated from Foxp3-GFP mice were differentiated into Treg in the presence or absence of Aza (5 μM), and equal numbers of Foxp3-GFP+ cells were harvested after 5 days of culture by FACS. The DNA was bisulfite converted, after which the TSDR region was PCR amplified and cloned, and the sequences analyzed for methylated CpG sites. Dramatic differences were evident between cells induced in the presence and absence of Aza. In the presence of Aza, the TSDR region was about 80% demethylated. In contrast, without Aza, the TSDR was only minimally demethylated (about 5%) (Fig. 7D). These methylation differences could have consequences in terms of Treg stability.
Since the Treg induced in the presence of Aza displayed a demethylated TSDR region, the effects of exposing the Treg population induced in the presence or absence of Aza to inflammatory cytokines that are known to destabilize Treg were measured (25, 42). The two Treg populations were harvested, and Foxp3 expression was determined following exposure for 3 days to IL-2 or IL-12 (Th1 conditions) or to IL-6 and TGF-β (Th17 conditions). In agreement with previous reports (43), IL-2 alone under nonstimulating conditions did not cause a change of Foxp3 expression. However, exposure to IL-12 for 3 days resulted in loss of Foxp3 expression in around 40% of cells. In contrast, the Treg induced in the presence of Aza lost only 20% of their Foxp3 expression after exposure to IL-12. Similar differences but with lesser magnitude were observed when the two populations were exposed to Th17 conditions (IL-6 and TGF-β). In those experiments, control-induced Treg lost around 25% of their Foxp3 expression, whereas Aza-induced Treg lost around 12% (Fig. 7E). In conclusion, Treg induced in vitro in the presence of Aza had TSDR that was demethylated, and such cells were more stable in the presence of inflammatory cytokines (IL-12 or IL-6) than were Treg induced without Aza.
To evaluate whether enhanced Treg stability may lead to enhanced Treg function, experiments were done to measure functional differences in Treg induced in vitro in the presence or absence of Aza. For these experiments, naive CD4 T cells isolated from Foxp3-GFP mice were used. The Foxp3-GFP+ cells were harvested 5 days after culture and subjected to FACS, and in vitro suppression assays were performed. Equal numbers of Foxp3+ T cells were cultured at different ratios, with naive responders being stimulated with anti-CD3/CD28 Abs. The results indicate that Treg differentiated in the presence of Aza showed more than twofold higher suppressive activity than control Treg (Fig. 8A and andB).B). In separate experiments, Treg were differentiated in the presence or absence of Aza (5 μM) to yield similar frequencies of Foxp3+ CD4 T cells between the two groups (high concentration of TGF-β) (Fig. 8C). The expression levels of ROS and the activation markers were compared. The Treg generated in the presence of Aza displayed around 1.3- to 1.8-fold increases in the expression of CD25, GITR, FR4, OX40, and ROS compared to the expression levels in control Treg (Fig. 8D and andE).E). In conclusion, exposure to Aza during Treg induction resulted in enhanced Treg suppressive function that could be partly explained by enhanced activation markers and ROS production in vitro.
Ocular infection with HSV sets off an inflammatory cytokine reaction in the cornea that leads to both virus clearance and chronic lesions that are orchestrated by CD4 T cells (4, 36). Approaches that enhance the function of Treg cells and dampen effector T cells can be effective to limit SK lesion severity (7,–10). In this report, we have explored the novel approach of inhibiting DNA methyltransferase activity using 5-azacytidine (cytosine analog) to limit HSV-induced ocular lesions. We show that therapy begun after infection, when virus was no longer actively replicating, resulted in a pronounced reduction in lesion severity, with markedly diminished numbers of inflammatory T cells and nonlymphoid inflammatory cells, along with reduced levels of cytokine mediators. The remaining inflammatory reactions had a change in the ratio of CD4 Foxp3+ Treg to effector Th1 CD4 T cells in ocular lesions, with Treg becoming predominant over the effectors. We also show that a consequence of Aza therapy was an increased suppressive activity of Treg, an effect which correlated with their increased expression of ROS. Hence, treatment with azacytidine during early stages of lesion development represents an effective and novel therapy for a lesion that is a common cause of human blindness (44).
Our results clearly showed markedly reduced lesions in response to HSV-1 infection in Aza-treated animals. In the model we used, SK lesions are immunopathological and are orchestrated mainly by IFN-γ-producing CD4 T cells. However, the tissue damage is mediated largely by neutrophils and, to a lesser extent, macrophages, which are recruited to the corneal site of inflammation by signals generated by the T cells (45, 46). In consequence, the inhibitory effects of Aza might be directed against multiple cell types in the SK response. In fact, some reports have indicated that Aza therapy can inhibit the generation of neutrophils (47) and proinflammatory (M1) macrophages (48), but we argue that the anti-inflammatory effects of Aza in the SK system may be explained mainly by its effects on T cells, particularly Treg. Thus, whereas all cell types were reduced in number in Aza-treated animals, there was a differential effect on Th1 effectors and Treg. In fact, in treated animals, the ratio of Treg to Th1 cells was increased substantially in corneal lesions (a change from 1:7 to 2:1), and similar but less dramatic changes in ratio occurred in the blood and DLN. Our results indicate that the ratio change may be more a consequence of direct effects on Treg than on T effectors. In fact, our working hypothesis is that Aza serves to stabilize, expand, or change the regulatory potency of Treg and that this acts to inhibit the function or perhaps transport of effectors to the corneal site of inflammation. Support of these ideas came from the observation that antigen-specific effectors were reduced in number in the DLN of Aza-treated animals, an effect that is likely the consequence of enhanced Treg function. Thus, we observed that Treg induced in the presence of Aza had significantly enhanced suppressive activity in vitro compared to that of cells from control animals.
With regard to why the Treg from Aza-treated animals were more suppressive than Treg from control animals, we could show that the levels of activation markers like CD25, GITR, OX40, and FR4, along with ROS, were significantly increased as a consequence of Aza therapy. A possible involvement of ROS activity in Treg function was noted in models of autoimmune arthritis and colitis, where inhibition of ROS-producing enzyme system components like NCF-1 and NOX2 led to loss of Treg suppressor function, enhancement of effector responses, and aggravation of inflammatory lesions (49,–51). Conceivably, the increase in ROS expression by Treg makes them more inhibitory against T effectors by inducing T cell death (52, 53). However, since Aza induces DNA demethylation across several genes, leading to their increased transcriptional activity, whether or not changed ROS expression is the most critical event that explains why Treg after Aza treatment were more effective in controlling the inflammatory reactions in the SK system needs further study. One line of studies we are pursuing to determine whether the promoter or the intron regions of Treg-associated genes like those encoding CD25, GITR, NCF-1, and NOX-2 might be hypomethylated upon Aza treatment, leading to their increased gene expression.
An alternative potential explanation for increased Treg representation over Th1 effectors in Aza-treated animals could be that Aza might render Treg resistant to the destabilization effects of proinflammatory cytokines that are highly expressed at the lesion and DLN sites (36, 54). Although this destabilization phenomenon was not evaluated in vivo, we could show that Treg induced in vitro in the presence of Aza, but not control Treg, displayed enhanced stability when exposed to the proinflammatory cytokines IL-12 and IL-6. The increased stability was likely explained by epigenetic differences caused by Aza therapy to inhibit DNA methyltransferase that were induced as downstream signaling events of proinflammatory cytokines. Such events result in methylation of the TSDR and only transient Foxp3 expression (25, 55). Evidence for epigenetic differences in the TSDR regions of Treg generated in the presence and absence of Aza was shown by in vitro studies. Thus, Treg generated in the presence of Aza had a demethylated TSDR region and showed stability when exposed to proinflammatory cytokines, unlike the non-Aza-exposed Treg, which had a methylated TSDR and lost Foxp3 expression in the presence of proinflammatory cytokines.
The final line of evidence implicating Treg as a critical cell type affected by Aza therapy came from the observation that the anti-inflammatory effects of Aza were blunted if Treg were depleted from animals prior to infection and subsequent Aza therapy. This observation also makes it unlikely that Aza acts to cause suppression of lesions via direct inhibitory effects on effector T cells or on nonlymphoid inflammatory cells like neutrophils. Supporting this notion, no inhibitory effects of Aza on effectors were observed in vitro, and in fact, when Aza was present during in vitro induction, both Treg and Th1 cells were increased in frequency. This observation that Aza therapy did not limit the lesion severity and effector responses when Treg were depleted came as a surprise, since the anti-CD25 MAb depletion procedure was only around 50% effective at depleting Treg. However, the depletion procedure is known to preferentially deplete Treg with high expression of the IL-2 receptor (CD25) (56, 57). Moreover, the CD25high population likely includes the antigen-specific Treg involved in regulating the effectors involved in SK. In fact, in prior studies, we had shown that CD25 depletion using anti-CD25 Ab results in enhanced effector function, along with more severe lesions of SK (58). Moreover, some studies have shown that CD25high Treg are in fact the precursors of antigen-specific Treg (59,–61), but we lacked the necessary reagents to formally demonstrate the antigen-specific Treg in our system. Nevertheless, Treg without HSV antigen specificity can also express modulatory effects in the SK system (7), although their CD25 expression level has not been evaluated. Overall, we take our observations to indicate that Aza therapy acts to stabilize and increase the regulatory function of Treg, an effect which likely acts in lymphoid tissue, as well as at the corneal inflammatory site, to limit the magnitude of effector T cell responses.
In conclusion, our results are consistent with the observation that inhibiting DNA methyltransferase activity through the use of azacytidine plays a role in influencing the expression of SK lesions. The mechanisms proposed to explain the outcome were multiple and involved a change in the balance between effector and regulatory T cells. We anticipate that inhibiting DNA methyltransferase could represent a useful approach to control an important cause of human blindness.
Female C57BL/6 mice and congenic Thy1.1+ mice were purchased from Harlan Sprague-Dawley, Inc. (Indianapolis, IN), Foxp3-GFP (C57BL/6 background) mice were a kind gift from M. Oukka (Brigham and Women's Hospital, Harvard Medical School), and BALB/c DO11.10 RAG2−/− mice were purchased from Taconic. Mice were kept in a pathogen-free facility where food, water, bedding, and instruments were autoclaved. All animals were housed in American Association of Laboratory Animal Care-approved facilities at the University of Tennessee, Knoxville, TN. All investigations followed the guidelines of the Institutional Animal Care and Use Committee and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. HSV-l strain RE was used in all procedures. Virus was grown in Vero cell monolayers (American Type Culture Collection, Manassas, VA), titrated, and stored in aliquots at −80°C until used.
Corneal infections of C57BL/6 and Foxp3-GFP mice were conducted under deep anesthesia induced by intraperitoneal injection of tribromoethanol (Avertin). Mice were scarified on the cornea with a 27-gauge needle, and a 3-μl drop containing 1 × 104 PFU of HSV was applied to the eye. The eyes were examined on different days postinfection (dpi) with a slit-lamp biomicroscope (Kowa Company, Nagoya, Japan), and the clinical severity of keratitis for individually scored mice was recorded as previously described (62). Briefly, the scoring system was as follows: 0, normal cornea; 1, mild corneal haze; 2, moderate corneal opacity or scarring; 3, severe corneal opacity but iris visible; 4, opaque cornea and corneal ulcer; 5, corneal rupture and necrotizing keratitis.
5-Azacytidine (MP Biomedicals) was dissolved in PBS and administered intraperitoneally at 2 mg/kg of body weight from day 5 postinfection until day 14 after infection. The control group either received an equal volume of PBS or was left untreated. The dose of Aza was chosen based on our preliminary studies (data not shown) and previous reports (63). Most of the experiments were repeated at least three times.
Eyes from control and Aza-treated mice were extirpated on day 15 p.i. and snap-frozen in optimum cutting temperature (OCT) compound (Miles, Elkhart, IN). Six-micrometer-thick sections were cut and air dried in a desiccation box. Staining was performed with hematoxylin and eosin (Richard Allen Scientific, Kalamazoo, MI).
At day 15 p.i., corneas were excised, pooled groupwise, and digested with Liberase (Roche Diagnostics Corporation, Indianapolis, IN) for 45 min at 37°C in a humidified atmosphere of 5% CO2. After incubation, the corneas were disrupted by grinding with a syringe plunger on a cell strainer, and a single-cell suspension was made in complete RPMI 1640 medium. The single-cell suspensions obtained from corneal samples were stained for different cell surface molecules for fluorescence-activated cell sorting (FACS) analyses. Draining cervical lymph nodes (DLN) were obtained from mice sacrificed at 15 days postinfection, and single-cell suspensions were used. Blood samples were collected at intervals from Aza-treated or control C57BL/6 Foxp3-GFP mice (HSV infected) to record the percentage of CD4+ T cells that were Foxp3 positive. All steps were performed at 4°C. Briefly, cells were stained with the respective fluorochrome-labeled cell surface antibodies in FACS buffer for 30 min and then stained for intracellular antibodies. Finally, the cells were washed three times with FACS buffer and resuspended in 1% paraformaldehyde. The stained samples were acquired with a FACS LSR II (BD Biosciences, San Jose, CA), and the data were analyzed using FlowJo software (Tree Star, Inc., Ashland, OR).
To determine the number of IFN-γ-producing T cells, intracellular cytokine staining was performed. In brief, corneal cells were stimulated either with phorbol myristate acetate (PMA) (50 ng) and ionomycin (500 ng) for 4 h in the presence of brefeldin A (10 μg/ml) or with UV-inactivated HSV-1 strain RE (multiplicity of infection of 1) overnight, followed by 5 h of culture with brefeldin A (10 μg/ml) in U-bottom 96-well plates (6). After this period, LIVE/DEAD staining was performed, followed by cell surface and intracellular cytokine staining using a Foxp3 intracellular staining kit (eBioscience) in accordance with the manufacturer's recommendations.
CD4 (RM4-5), CD45 (53-6.7), CD11b (M1/70), Ly6G (1A8), F4/80 (BM8), IFN-γ (XMG1.2), CD103 (M290), CD25 (PC61 and 7D4), GITR (DTA-1), FR4 (eBio12A5), CD44 (IM7), OX40 (OX-86), annexin-V, Foxp3 (FJK-16S), anti-CD3 Ab (145-2C11), anti-CD28 Ab (37.51), and GolgiPlug (brefeldin A) were from either eBioscience or BD Biosciences. PMA and ionomycin were from Sigma. CellTrace violet (CTV), LIVE/DEAD fixable violet dead-cell stain kit, and CM-H2DCFDA (6-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate, acetyl ester) were from Life Technologies. Recombinant IL-2 (rIL-2), rIL-12, rIL-6, and rTGF-β were from R&D Systems.
At day 15 after ocular infection with HSV-1, the corneas were isolated and four corneas were pooled per sample/group. Regulatory T cells and effector cells were subjected to FACS using Foxp3-GFP mice. Total mRNA from corneal and sorted T cell populations was isolated using the mirVana miRNA isolation kit (Ambion). cDNA was made with 500 ng of RNA (corneal samples) and entire RNA (isolated T cells) by using oligo(dT) primer and the ImProm-II reverse transcription system (Promega). TaqMan gene expression assays for cytokines (IL-10, TGF-β, IL-1β, TNF-α, IL-6, and IL-12), chemokines (CCL3, CXCL1, CCL2, and MMP1), and NADPH oxidase components (NCF-1 and NOX2) were purchased from Applied Biosystems and quantified using the 7500 Fast real-time PCR system (Applied Biosystems). The expression levels of different molecules were normalized to that of β-actin using the cycle threshold (ΔCT) method. Relative expression between control and experimental groups was calculated using the formula 2−ΔΔCT × 1,000.
C57BL/6 mice were given intraperitoneal injections of 500 μg of anti-CD25 MAb (clone PC61 rIgG1; BioXcell, West Lebanon, NH) or control rat IgG1 (BioXcell) on the same day of infection (day 0). The Treg depletion efficiency was quantified by measuring the percentage of Foxp3+ CD4 T cells at 15 days postinfection.
CD4+ T cells (total or naive) were purified from single-cell suspensions of pooled DLN and spleens from HSV-infected or naive Foxp3-GFP and Thy1.1+ B6 (H-2b) mice using a mouse total or naive CD4+ T cell isolation kit according to the manufacturer's instructions (Miltenyi Biotec, Auburn, CA). At least 90% purity was achieved. For methylation studies and suppression assays, Treg cultures were sorted based on Foxp3-GFP using FACS to achieve high purity.
The Treg suppression assay was done as previously described (9). Briefly, Foxp3-GFP mice infected with HSV-1 were divided into multiple groups. Mice in one group were injected with Aza on day 5 p.i., and control groups were injected with PBS. At day 15 p.i., single-cell suspensions from DLN and spleens were prepared, and CD4+ Foxp3+ T cells were sorted on a FACSAria cell sorter to 99% purity. To measure the suppressor function of Treg differentiated in vitro, naive CD4 T cells from Foxp3-GFP mice were differentiated to Treg in the presence or absence of Aza and Foxp3-GFP+ cells subjected to FACS. CD4+ Foxp3+ T cells were then cultured with anti-CD3 (1 μg/well) and anti-CD28 (0.5 μg/well) antibodies and CTV-labeled naive CD4+ Thy1.1 responder cells (purified by a Miltenyi Biotec kit) in a 96-well round-bottom plate. The suppressive capacity of Treg was measured by coculturing Treg and conventional T cells (Tconv) at different ratios (Treg/Tconv, 1:1 to 1:16). After 3 days of incubation, the extent of CTV dilution in Thy1.1 CD4+ cells was measured by flow cytometry. The percentage of suppression by Treg was calculated by using the formula 100 − [(frequency of cells proliferated at a particular ratio of Treg to effector T cells)/(frequency of cells proliferated in the absence of Treg)]100.
Splenocytes isolated from DO11.10 RAG2−/− or Foxp3-GFP mice were used as a precursor population for the induction of Foxp3+ in CD4+ T cells as previously described (8). Briefly, after red blood cell (RBC) lysis and several washings, 1 × 106 splenocytes were cultured in 1 ml RPMI medium containing rIL-2 (100 U/ml) and TGF-β (1 to 5 ng/ml) in the presence or absence of various concentrations of Aza (1 to 15 μM) with plate-bound anti-CD3/CD28 Abs (1 μg/ml) for 5 days at 37°C and 5% CO2 in an incubator. After 5 days, samples were characterized for Foxp3 intracellular staining (eBioscience staining kit) or GFP expression (Foxp3-GFP mice) by flow cytometry. Treg were either sorted (TSDR methylation analysis) or cultured in a 96-well round-bottom plate in the presence of IL-2 (100 U/ml), IL-12 (5 ng/ml), or IL-6 (25 ng/ml) plus TGF-β (1 ng/ml) for 3 days, followed by flow cytometry analysis of live CD4+ Foxp3+ cells.
For Th1 differentiation, splenocytes from DO11.10 RAG2−/− mice were stimulated with plate-bound anti-CD3/CD28 Abs (1 μg/ml) in the presence of recombinant mouse IL-12 (5 to 10 ng/ml) and anti-IL-4 Ab (10 μg/ml) and in the presence or absence of various concentrations of Aza (1 to 15 μM). After 5 days, samples were restimulated with PMA-ionomycin and analyzed for the production of IFN-γ by using an intracellular cytokine staining kit (BD Biosciences) and a flow cytometer.
Foxp3-GFP+ cells were subjected to FACS, and genomic DNA was isolated (Qiagen) and bisulfite converted with an EZ DNA Methylation-Direct kit according to the manufacturer's protocol (Zymo Research). The TSDR region (corresponding to conserved noncoding sequence 2 of the Foxp3 gene) was PCR amplified using primer sequences 5′-GGGTTTTTTTGGTATTTAAG-3′ (forward) and 5′-CCTAAACTTAACCAAATTTT-3′ (reverse). The PCR products were subcloned into pGEM-T Easy vectors (Promega) and transformed into bacterial clones. Plasmid DNA samples from each bacterial colony were sequenced separately at the UTK core facility (at least 10 sequences per sample).
Single-cell suspensions from DLN from both Aza-treated and control HSV-infected C57BL/6 mice or in vitro-differentiated Treg were incubated with 1 μM CM-H2DCFDA for 30 min at 37°C, followed by washing with PBS and surface staining for live CD4+ CD25+ cells. Oxidation of dye was detected by fluorescein isothiocyanate (FITC) fluorescence.
Statistical significance was determined by Student's t test unless otherwise specified. A P value of <0.05 was regarded as a significant difference between groups. GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA) was used for statistical analysis.
We thank Naveen Rajasagi for assistance during preparation of the manuscript. We also thank Benjamin A. Youngblood from St. Jude Children's Hospital for assistance in TSDR methylation analysis.
This study was supported by National Institutes of Health grant number EY 005093 and National Institute of Allergy and Infectious Diseases grant number AI 063365.