Altered bone marrow dendritic cell cytokine production to toll-like receptor and CD40 ligation during chronic feline immunodeficiency virus infection

Abstract

Impaired dendritic cell (DC) function is thought to be central to human immunodeficiency virus-associated immunodeficiency. In this study, we examined the effect of chronic feline immunodeficiency virus (FIV) infection on DC cytokine production in response to microbial and T-cell stimulation. Cytokine production after either Toll-like receptor (TLR) or CD40 ligation in bone marrow-derived DCs (BM-DCs) was measured in naïve and chronically FIV-infected cats. The BM-DCs were stimulated with ligands to TLR-2, -3, -4, -7 and -9 or cocultured with 3T3 cells expressing feline CD40 ligand. Ligation of TLR-4 and TLR-9 in BM-DCs from infected cats resulted in a significant decrease in the ratio of interleukin-12 (IL-12) to IL-10. Conversely, TLR-7 ligation produced a significant increase in the IL-12 : IL-10 ratio in BM-DCs from infected cats. No difference was noted for TLR-3 ligation. RNA expression levels of TLR-2, -3, -4, -7 and -9 were not significantly altered by FIV infection. CD40 ligation significantly elevated both IL-10 and IL-12 messenger RNA production but did not alter the IL-12 : IL-10 ratio. Chronic FIV infection alters the ratio of immunoregulatory cytokines produced by BM-DCs in response to certain pathogen-derived signals, which is probably relevant to the increased risk of opportunistic infections seen in lentiviral infection.

Introduction

During a successful immune response, a pathogen must be recognized and appropriate cytokines must be produced by antigen-presenting cells, including dendritic cells (DCs). Toll-like receptors (TLRs) recognize conserved molecules on pathogens and play a central role in this process inducing the maturation and activation of DCs.¹ Mature, activated DCs then stimulate T cells, providing a link between the innate and adaptive immune responses. Thirteen TLRs have been described in mammalian cells in recent years.² The TLRs 1, 2, 4, 5 and 6 are expressed on the cell surface while TLRs 3, 7, 8 and 9 are located within endosomes, although cell type and maturation state can change the expression and localization of TLRs.² These TLRs recognize pathogen-associated molecular patterns and induce a complex signalling cascade within the cells that results in the production of cytokines. Unique among the known TLRs, the signalling cascade induced by TLR-3 ligation does not use the Toll/interleukin-1 (IL-1) receptor domain-containing adaptor MyD88.

Many pathogens, including viruses, have evolved mechanisms that evade or use TLR-driven host defence mechanisms to avoid immune clearance. Respiratory syncytial virus induces TLR-4 signalling, resulting in destructive inflammation of the respiratory epithelium.³ The retrovirus mouse mammary tumour virus uses TLR-4-induced IL-10 production to escape the immune response.⁴ Human immunodeficiency virus (HIV) exploits TLR-2 ligation by concurrent mycobacterial infection to up-regulate viral replication.⁵ Feline immunodeficiency virus (FIV) infection has been shown to result in altered TLR expression on lymphocyte subsets, although the functional significance of this finding is not yet clear.⁶

Dendritic cells predominate on mucosal surfaces and are one of the first targets of lentiviruses such as HIV, simian immunodeficiency virus, and FIV.^7,8 Lentiviral infection has been shown to alter DC growth, number, maturation and function in favour of viral proliferation and spread, although these changes are not fully understood and conflicting evidence exists⁹. Lentiviruses probably induce changes in DCs both through direct effects of viral infection and through virus-induced alterations in cytokine production.¹⁰

Feline immunodeficiency virus is a lentivirus that produces disease strikingly similar to acquired immunodeficiency syndrome (AIDS) with clinical phases analogous to those of HIV.^11,12 The FIV provides a model of HIV infection in which the components of the immune response to a lentivirus and the resulting changes induced by such a virus can be studied. We used feline bone marrow-derived DCs (BM-DC) expanded with granulocyte–monocyte colony-stimulating factor (GM-CSF) to investigate TLR expression and signalling in cats chronically infected with FIV. Bone marrow mononuclear cells grown with GM-CSF have been shown to be myeloid dendritic cells (mDCs),¹³ and bone marrow yields greater numbers of DCs with fewer macrophages than peripheral blood mononuclear cells (our unpublished data).¹⁴ Bone marrow-derived mDCs have high major histocompatibility complex (MHC) class II and costimulatory molecule expression and are efficient at initiating primary immune responses. Myeloid DCs produce IL-12 and direct cell-mediated immune responses, both of which are altered in HIV-infected individuals.^15,34

We also investigated the ability of BM-DCs from FIV-infected cats to respond to T-cell-derived signals necessary for amplification of the immune response. Interactions between the CD40 surface receptor on DCs and CD40 ligand (CD40L, also known as CD154) on CD4⁺ T cells have been shown to be crucial for the induction of a cell-mediated immune response. Viruses such as measles virus have been found to interfere with CD40-induced DC maturation as well as TLR signalling.^16,17 Conflicting research exists relative to HIV. Kawamura et al. have reported that in vitro HIV-1 infection impairs the capacity of DCs to stimulate CD4⁺ T cells but that HIV-1-exposed DCs secrete increased amounts of IL-12 p70 after stimulation with soluble CD40L.¹⁸ Alternatively, Smed-Sorensen et al. have shown that HIV infection impairs DC IL-12 p70 production in response to CD40L.¹⁹

In the present work, we demonstrate that FIV infection alters the ratio of pro- and anti-inflammatory cytokine production by BM-DCs in response to certain TLR ligands. Feline immunodeficiency virus also affects the levels of both IL-10 and IL-12 produced by BM-DCs in response to CD40–CD40L interactions, but the overall ratio of the two cytokines remains unchanged keeping the pro- and anti-inflammatory cytokine balance intact.

Materials and methods

Animals

Twenty-one naïve specific-pathogen-free cats and 15 specific-pathogen-free cats which had been infected with the molecular clone FIV-C36 for 12–18 months were used in this study. All animals were housed and treated in accordance with Colorado State University Animal Care and Use Committee approved protocols.

Cells

Bone marrow was collected from the proximal humerus of specific-pathogen-free and chronic FIV-infected cats. Bone marrow mononuclear cells were isolated using Ficoll–histopaque gradient centrifugation (Sigma-Aldrich, St Louis, MO) and stored in liquid nitrogen until use. The bone marrow mononuclear cells were cultured at an initial concentration of 5 × 10⁶/ml in six-well plates at 37° in 5% CO₂ for 10 days in medium consisting of RPMI-1640 media with Glutamax (Sigma-Aldrich) with 15% fetal bovine serum (Atlanta Biologicals, Norcross, GA), penicillin (100 U/ml; Invitrogen, Carlsbad, CA), streptomycin (100 μg/ml; Sigma-Aldrich) and 2-mercaptoethanol (50 μm; Sigma-Aldrich). Fifty per cent of the medium was changed and 100 ng/ml recombinant feline (rf) GM-CSF (R&D Systems, Minneapolis, MN) was added to the media every other day. After 10 days in culture, cells were stained for MHC class II (Tu39, BD PharMingen, San Jose, CA) and CD11c (CA11.6A1, Serotec, Raleigh, NC) and sorted using immunomagnetic beads (Miltenyi Biotec, Auburn, CA). Purity of the samples was confirmed by flow cytometry to be consistently greater than 95%.

TLR expression

Purified BM-DCs were lysed with Trizol (Gibco, Grand Island, NY), and the total RNA was extracted. Complementary DNA (cDNA) was made using the iScript cDNA kit (Bio-Rad, Hercules, CA). Real-time quantitative polymerase chain reaction (PCR) was performed to quantify RNA levels of TLRs 2–9 as previously described.⁶

Cytokine production via TLR stimulation

Purified BM-DCs were allowed to recover at 37° overnight in fresh LBT medium with rfGM-CSF (R&D Systems; 100 ng/ml) after sorting. The cells were then stimulated for 6 hr with one of the following TLR ligands: poly(I : C) (25 μg/ml; InvivoGen, San Diego, CA), ultrapure lipopolysaccharide (1 μg/ml; InvivoGen), Loxoribine (100 μm; InvivoGen), Pam2CSK4 (75 ng/ml; InvivoGen), or Escherichia coli single-stranded DNA (10 μg/ml; InvivoGen). All TLR ligands have been tested by the manufacturer to exclude non-specific TLR signalling. After 6 hr, the cells were lysed with Trizol (Gibco), and the RNA was extracted. The cDNA was made using the iScript cDNA synthesis kit (Bio-Rad). Real-time PCR for cytokines IL-12, IL-10, IL-6, tumour necrosis factor-α (TNF-α), and interferon-α (IFN-α) was performed as previously described.²⁰ The messenger RNA (mRNA) levels of cytokines from stimulated cells were compared with those from unstimulated cells in a relative quantitative assay using 18s RNA to normalize for the quantity of input RNA and the 2^ΔΔCT method was used to determine the relative expression of the genes of interest.

Cytokine production via CD40–CD40L interaction

3T3 cells expressing feline CD40L (3T3.CD40L) or a control 3T3 cell line not expressing CD40L, both generously provided by Margaret Hosie (University of Glasgow),²¹ were grown to approximately 80% confluency in a 24-well plate at 37° and 5% CO₂. The 3T3 cells were cultured in Dulbecco’s modified Eagle’s medium (Gibco) with 10% fetal bovine serum (Atlanta Biologicals), 2% glutamine (Invitrogen), and 1% penicillin–streptomycin (Invitrogen). For cell selection purposes, 400 μg/ml G418 sulphate (InvivoGen) was added to the media for CD40L-expressing 3T3 cells. Purified BM-DCs from seven naïve cats and six FIV-C-infected cats were allowed to rest overnight in fresh LBT medium with rfGM-CSF (R&D Systems; 100 ng/ml). The medium was removed from the 3T3 cells which did or did not express CD40L, and 1·5 × 10⁵ BM-DCs in 1 ml LBT medium was added per well. After 6 hr of culture, the cells were lysed with Trizol (Gibco), and the RNA was extracted according to the manufacturer’s instructions. Cytokine mRNA levels were measured as above. Cytokine RNA production in BM-DCs exposed to 3T3.CD40L cells was expressed as an increase over that of BM-DCs exposed to 3T3 cells.

Statistical analysis

The Kolmogorov–Smirnov test was performed to determine whether the data were normally distributed. Student’s t-test was performed and P < 0·05 was considered significant.

Results

High cell purity obtained by sorting

Flow cytometric analysis of magnetic bead and high-speed flow cytometry sorting methods based on MHC class II and CD11c expression showed > 90% (and generally > 95%) purity for all samples before stimulation and/or analysis (data not shown). The purified population of cells stimulates a robust mixed leucocyte response when derived from naïve animals (data not shown).

No difference in TLR mRNA expression in feline BM-DCs

The BM-DCs from six infected and eight naïve cats were analysed for expression of TLR-1, -2, -3, -4, -5, -7, -8 and -9 mRNA. Feline BM-DCs were found to express all of the TLRs but had low expression of TLR-3, -5 and -9. There was no significant difference in expression of the measured TLRs between infected and naïve cats (Fig. 1) although there was a trend for increased expression of TLR-2 in infected BM-DCs (P= 0·069, Student’s t-test).

Figure 1.

Toll-like receptor (TLR) messenger RNA (mRNA) expression is not altered in bone marrow-derived dendritic cells (BM-DCs) from cats infected with feline immunodeficiency virus (FIV). The BM-DCs were cultured for 10 days in the presence of recombinant feline granulocyte–macrophage colony-stimulating factor (100 ng/ml) and then purified for CD11c and major histocompatibility complex class II expression. Purified BM-DCs were lysed with Trizol (Gibco), and the RNA was extracted. Real-time polymerase chain reaction was performed to quantify RNA levels of TLRs 2, 3, 4, 5, 7 and 9. Results are shown as the mean and SD of cells from eight naïve cats and six FIV-infected cats. There were no statistically significant differences in expression levels as determined by Student’s t-test.

IL-12 : IL-10 ratio is altered with FIV infection

The TLR ligands for TLR-2, -3, -4, -7 and -9 were used to test cytokine mRNA production in BM-DCs from a minimum of five naïve and five FIV-infected cats. Regardless of the TLR expression level, all five ligands induced the expected cytokine responses in BM-DCs from both FIV-infected cats and naïve cats. No significant difference between the two groups was seen for any of the individual cytokines, i.e. IL-6, IL-10, IL-12, TNF-α and IFN-α (data not shown). However, looking at individual animals, the IL-12 : IL-10 ratio was significantly different between the two groups for TLR-4, -7 and -9 ligation. TLR-4 and TLR-9 ligation resulted in significantly lower IL-12 : IL-10 ratios in FIV-infected cats compared to naïve cats (P < 0·05, Student’s t-test; Fig. 2). TLR-2 showed a trend for the same change but this was not statistically significant. TLR-7 showed the opposite effect in that infected cats produced a significantly higher IL-12 : IL-10 ratio than naïve cats (P < 0·05, Student’s t-test; Fig. 2). TLR-3 ligation did not result in any differences in the IL-12 : IL-10 ratio between naïve and infected animals.

Figure 2.

There were no statistical differences in the relative induction of interleukin-10 (IL-10) or IL-12 RNA (a and b) when dendritic cells (DCs) were stimulated with any of the Toll-like receptor (TLR) ligands but the average individual cat IL-12 : IL-10 ratio (c) was significantly decreased with TLR-4 and TLR-9 ligation and increased with TLR-7 ligation in bone marrow derived (BM-) DCs from FIV-infected cats (asterisk indicates P < 0·05 by Student’s t-test). The BM-DCs were cultured for 10 days in the presence of recombinant feline granulocyte–macrophage colony-stimulating factor (100 ng/ml) and then purified for CD11c and major histocompatibility complex class II expression. Purified BM-DCs were cultured for 6 hr with a TLR ligand to TLR-2, -3, -4, -7 or -9. Real-time polymerase chain reaction for the cytokines IL-12, IL-10 was performed. The 18s RNA was used to normalize input RNA, and the 2^−ΔΔCT method was used to determine the relative expression of the genes of interest.

CD40L-induced cytokine production is altered in FIV infection

Cytokine mRNA was measured for IL-6, IL-10, IL-12, TNF-α, and IFN-α in BM-DCs cocultured with 3T3 cells that did or did not express feline CD40L. Increases were seen in IL-6, IL-10, IL-12 and TNF-α during co-culture of the BM-DCs with the CD40L-expressing cells as compared to BM-DCs cocultured with native 3T3 cells (Fig. 3a). The CD40L-induced production of both IL-10 and IL-12 were significantly increased in BM-DCs from FIV-infected cats compared with naïve cells (P= 0·038 and P= 0·036 respectively, Student’s t-test; Fig. 3a). Despite the individual increases in IL-12 and IL-10, the resultant ratio of induced IL-12 : IL-10 was unchanged between naïve and infected BM-DCs (P= 0·94, Student’s t-test; Fig 3b). No cytokine production was seen when 3T3 cells were grown in isolation.

Figure 3.

Bone marrow-derived dendritic cells (BM-DCs) from cats infected with feline immunodeficiency virus (FIV) produce relatively more interleukin-10 (IL-10) and IL-12 when exposed to CD40 ligand (CD40L) than BM-DCs from naïve cats yet the ratio of IL-12 and IL-10 induction remains unchanged. Purified BM-DCs from seven naïve cats and six FIV-C-infected cats were cultured for 6 hr with 3T3 cells, which did or did not express CD40L. After 6 hr, the cells were lysed, and the RNA was extracted. Cytokine messenger RNA (mRNA) was measured for IL-6, IL-10, IL-12, tumour necrosis factor-α (TNF-α), and interferon-α (IFN-α) by real-time polymerase chan reaction. Cytokine mRNA production in BM-DCs exposed to 3T3 cells expressing CD40L is expressed as an increase over that of BM-DCs exposed to 3T3 cells that do not express CD40L. (a) Exposure to CD40L induced the expression of IL-6, IL-10, IL-12, TNF-α and IFN-α mRNA in feline BM-DCs. The BM-DCs from FIV-infected cats produce more IL-10 and IL-12 mRNA than cells from naïve cats (P < 0·05, Students t-test). (b) No significant change was noted in the IL-12 : IL-10 ratio between cells from naïve and FIV-infected cats (P > 0·05, Student’s t-test).

Discussion

Lentiviruses subvert the normal immune response to establish long-term infections despite or because of the efforts of DCs and other immune effector cells. Dendritic cells are thought to become dysfunctional during HIV infection, although the exact mechanism(s) of this dysfunction are not yet clearly defined. There is evidence that HIV and other pathogens interfere with the normal function of TLRs and use the TLRs to promote infection.^5,16,22 The ability of DCs to successfully interact with and stimulate T cells during lentiviral infection has also come into question. We sought to determine if chronic FIV infection results in altered DC cytokine production in response to TLR stimulation and/or the CD40–CD40L DC : T-cell interaction, thereby providing further insight into lentiviral affects on DC function.

Dendritic cell subsets have yet to be clearly defined in cats, yet our feline BM-DCs are CD11c⁺ CD11b⁺CD1a⁺ MHCII ^high, which is consistent with myeloid origin, and murine BM-DCs grown with GM-CSF have been shown to develop into mDCs.¹³ The retained expression of CD14 on feline DCs is consistent with data from other laboratories. ^14,23 Similar to human mDCs, we found that feline BM-DCs grown with GM-CSF express high levels of TLR-2 and TLR-8, moderate levels of TLR-1, -4 and -7, and low levels of TLR-3, -5 and -9.^24,25 We did not find a correlation between the amount of TLR RNA expression and the overall magnitude of the cytokine response to individual TLR ligands. Although it has been reported that TLR expression correlates with cytokine production, measurement of RNA expression levels of TLRs may not accurately predict their activity. Expression of TLR-4 and TLR-9 is low in immature human mDCs but the cells are able to respond strongly to lipopolysaccharide and specific CpG motifs, suggesting either an alternative pathway or high efficiency of the receptor.^24,26,27 There were no significant differences in TLR expression between BM-DCs from naïve or chronically FIV-infected cats. The high variability in TLR expression observed among individual cats has also been noted for human donors.²⁸

When cytokine production was measured, we found that BM-DCs from FIV-infected cats were able to produce similar increases in IL-12, IL-10, TNF-α, IL-6 and IFN-α mRNA as BM-DCs from naïve cats in response to individual stimulation of the five TLRs we tested (TLRs 2, 3, 4, 7 and 9). There was considerable variation in the overall magnitude of cytokine responses among individual animals, which is not surprising in an outbred population. Recent work indicates that people can be classified as low or high responders based on the magnitude of cytokine production in response to lipopolysaccharide stimulation.²⁹

To account for the variability in the overall magnitude of cytokine response between individual animals, we examined the balance of cytokines that each animal produced. We found that the ratio of IL-12 and IL-10 in individual cats was significantly altered in FIV-infected versus naïve control cats. In particular, simulation of TLR-4 and TLR-9 shifted the cytokine balance towards IL-10 in BM-DCs from FIV-infected cats whereas stimulation of TLR-7 shifted the balance towards IL-12 in the same animals. Concurrently, the MyD88-independent pathway stimulated by ligation of TLR-3 was unaffected. Early DC production of IL-12 is critical to initiating an effective cell-mediated immune response by stimulating T-cell or natural killer cell production of IFN-γ.³⁰ Increases in IL-10 production can interfere with DC activation of the adaptive immune response by decreasing IL-12 production,³⁰ DC surface MHC class II expression³¹ and antigen presentation,³¹ and by impairing DC maturation through inhibition of costimulatory molecule expression.³² Decreased costimulatory molecule expression and impaired DC maturation have been associated with the development of a type 2 cytokine response, regulatory T cells and tolerance.^33,34

Many pathogens have been shown to alter this early DC cytokine balance which may inhibit immune clearance. Myeloid DCs from patients infected with hepatitis C virus produce higher amounts of IL-10 in response to TLR-3 ligation with poly-IC.³⁵ Measles virus, which causes immunosuppression and enhanced susceptibility to microbial insults, has been found to suppress IL-12 synthesis through TLR-4.¹⁶ The Gram-negative bacterium Bordetella bronchiseptica, which establishes a persistent infection in the murine respiratory tract, expresses virulence factors which have been shown to decrease IL-12 production in BM-DCs and to induce a semi-mature phenotype in these same cells.³⁶ Recent work demonstrates that a soluble extract from the eggs of the helminth Schistosoma mansoni, a cause of the chronic and often debilitating disease schistosomiasis, inhibits conventional human monocyte-derived DC maturation and decreases IL-12, IL-6 and TNF-α production in response to TLR-3 and TLR-4 ligands.³⁷

A selective impairment in the cytokine response to the ligation of TLR-4 or TLR-9 could be instrumental in the pathogenesis of many of the opportunistic infections seen during AIDS. Secondary bacterial infections are the leading cause of death in HIV-infected patients. Gram-negative organisms such as Salmonella spp. and Pseudomonas aeuroginosa are commonly documented in patients with AIDS and are recognized by TLR-4.^38,39 Human herpes virus infections are associated with significant morbidity in HIV-infected individuals and the generation of an effective immune response to both herpes simplex virus and cytomegalovirus has been shown to be dependent on an intact DC TLR-9 signalling pathway.^40,41 TLR-4 has been shown to be critical in controlling other AIDS-defining illnesses, including disseminated mycobacterial infection and candidiasis.^42,43 Cats infected with FIV often succumb to disseminated bacterial infections, isolation of Candida albicans has been reported with more frequency in FIV-infected cats and rare feline cases of disseminated mycobacterial infection have been associated with FIV infection.^44,45 Further work using whole organisms would be required to more directly test this theory of altered DC-mediated pathogen response.

Interestingly, TLR-7 stimulation of DCs from FIV-infected cats resulted in the opposite pattern of cytokine expression favouring the production of the proinflammatory cytokine IL-12 over IL-10. Agonists of TLR-7 have been shown to preferentially induce IL-12 production from circulating murine CD11c⁺ mDCs leading to subsequent T-cell T helper type 1 cytokine polarization.⁴⁶ Measles virus infection of DCs has recently been shown to selectively impair TLR-4-mediated IL-12 production while resulting in the enhanced production of IL-12 after TLR-7 ligation.¹⁶ Binding of HIV single-stranded RNA to TLR-7 and TLR-8 on latently infected promonocytic cells results in the production and release of HIV virions despite the fact that it appears to have an anti-viral effect on acutely infected cells.⁴⁷ Signalling through TLR-7 may be preserved and even augmented in lentiviral infections if it plays a necessary role in regulating viral replication and latency. These results are of particular interest given the use of selective TLR-7 agonists in preclinical trials of several infectious diseases and argue that preferential TLR-7 stimulation may be of benefit in chronic lentiviral infections to both stimulate IL-12 production and reactivate latent viral reservoirs during treatment.

CD40–CD40L ligation induces up-regulated DC surface expression of MHC class I and II and the costimulatory molecules CD80 and CD86 and enhanced production of cytokines such as IL-12 and IL-15, which are important for T-cell growth and polarization.^48,49 We found that BM-DCs from FIV-infected and naïve cats induced equivalent IL-6, TNF-α, and IFN-α mRNA production in response to CD40–CD40L ligation. The production of both IL-10 and IL-12 was found to be significantly increased in BM-DCs from FIV-infected cats, which led to maintenance of the balance of IL-12 and IL-10 seen in naïve cats. The in vivo significance of a relative increase in the amount of IL-10 and IL-12 mRNA but maintenance of the same pro- and anti-inflammatory cytokine balance is as yet unknown. Zhang et al. have shown that coculture of DCs with HIV-exposed CD4⁺ T cells resulted in decreased DC IL-12 production, but the decrease was the result of decreased CD40L expression by the T cells, not of DC dysfunction.⁵⁰ Other work has shown that bulk culture of HIV-exposed DCs revealed no differences in CD40–CD40L-induced maturation status or IL-12 production, yet in these same experiments isolation of the small number of infected DCs showed that they were unable to produce IL-12p70.¹⁹ We are currently unable to specifically isolate FIV-infected DCs to evaluate them in a similar fashion. The exact mechanisms by which FIV and HIV affect the CD40–CD40L interaction and their in vivo significance remain to be defined.

The relative balance of DC IL-12 and IL-10 production can determine the outcome of an intracellular infection, and it appears that pathogens including FIV have evolved ways to exploit this balance. The selective alteration of TLR-induced DC cytokine responses may allow FIV to evade immune clearance, predispose the host to AIDS-defining secondary infections, and play a role in regulating viral latency. Feline immunodeficiency virus would not be the first virus known to induce both pro- and anti-inflammatory cytokines during the course of infection. Isolating the exact signalling modification(s) induced by FIV in these cells would shed light on the viral mechanisms of immune alteration and possibly help identify target areas of intervention for lentiviral infections. Further studies to identify the exact mechanism involved in alteration of the TLR-induced signalling pathways are ongoing.