Interfer’n with antibody responses

See allHide authors and affiliations

Science Immunology  21 Oct 2016:
Vol. 1, Issue 4, eaaj1836
DOI: 10.1126/sciimmunol.aaj1836


Type I interferon blocks the generation of neutralizing antibodies in response to chronic infection.

Chronic infection is often associated with poor induction of neutralizing antibody responses. Such is the case in humans infected with HIV, hepatitis B virus, and hepatitis C virus. This contrasts with acute infections by cytopathic viruses where there is usually robust induction of neutralizing antibodies (1). Multiple studies have attempted to determine how infection by low or noncytopathic, chronically infecting viruses thwarts B cell responses, and a range of potential mechanisms have been described, including low B cell precursor frequency, disruption of lymphoid tissue organization, misdirected CD4 T cell and nonneutralizing B cell responses, and killing of cells by cytotoxic T lymphocytes (CTLs) (1). However, it has been unclear whether there is a common thread tying these processes together or which processes are more important in suppressing antibody responses. This dissatisfying situation is transformed by three studies in this issue of Science Immunology that establish type I interferon (IFN-I) as a major factor induced by chronic viral infection that suppresses induction of the neutralizing antibody response (24). The breadth of investigation across the three studies and their complementary nature lead to a rich picture of the mechanisms involved, with each highlighting a distinct component of the IFN-I mode of action.

IFN-I’s, which include more than a dozen IFN-α family members and a single IFN-β, are induced rapidly upon viral infection of cells and act on the IFN-I receptor, also called interferon-α/β receptor (IFNAR), which signals via JAK/STAT pathways to induce hundreds of so-called IFN-stimulated genes (ISGs) (5). ISGs often have direct antiviral functions, but many are also immunomodulatory; whether the effects are positive or negative for adaptive immune responses is strongly influenced by context and timing (5). The IFNAR is widely expressed by both hematopoietic and nonhematopoietic cell types, allowing IFN-I to influence the properties of a diversity of cells.

Lymphocytic choriomeningitis virus (LCMV), a member of the Arenaviridae family of single-stranded RNA viruses, is a natural rodent pathogen. The virus has a single surface glycoprotein (GP) that is the target of neutralizing antibodies. All three studies use the clone 13 strain of LCMV that causes a noncytopathic, chronic infection in mice. Although there is early induction of a nonneutralizing LCMV-reactive antibody response, there is a 2- to 3-month delay before appearance of neutralizing antibodies (1). Many cell types are infected by LCMV, including myeloid cells, stromal cells, and lymphocytes. The three groups compared events after LCMV infection to those occurring after infection by the acute cytopathic virus—vesicular stomatitis virus (VSV)—that induces a prompt neutralizing antibody response (1). The studies took advantage of an established line of mice, KL25H, that express an LCMV GP-specific immunoglobulin (Ig) heavy chain from the endogenous Ig locus, along with newly generated light-chain transgenic mice. The use of KL25HL GP-specific B cells allowed each study to ensure that any response defect is not a consequence of deficiency in antigen-specific precursor B cells.

To gain insights into the earliest repressive factors, Fallet et al. (2) performed a careful side-by-side comparison of the KL25HL transgenic B cell response to LCMV and recombinant VSV expressing the LCMV GP. They found that GP-specific B cells rapidly give rise to antibody-secreting cells (ASCs) after both infections. However, in the LCMV-infected mice, the conversion to short-lived ASCs was so complete that the B cell response peaked within a few days and then decays away, whereas the response against VSV continued to rise for 2 to 3 weeks. The comparative analysis reveals that the period of B cell response repression coincided with the time of highest serum IFN-I levels, and antibody blocking of IFNAR had a remarkably restorative effect on the response. The effect of IFN-I is B cell–extrinsic, yet largely hematopoietic cell–intrinsic. Using cell type–specific ablation of IFNAR, they conclude that strong IFN-I induction after infection suppresses virus-specific B cell responses through effects on multiple cell types. Myeloid cells, dendritic cells, and T cells all participate, and the cytokines interleukin-10 (IL-10) and tumor necrosis factor (TNF) play a role. The suppressive effect of IFN-I involves these (and other) factors acting to “decimate” antigen-activated B cells through differentiation into short-lived ASCs, and this prevents cells taking on a germinal center (GC) or memory cell fate (Fig. 1).

Fig. 1 Pleiotropic actions of IFN-I in repressing virus-neutralizing antibody (Ab) responses.

LCMV infection induces IFN-I that promotes increased function of CTL; production of IL-10, TNF, and other cytokines; and, in LNs, recruitment of NO-producing inflammatory monocytes. LCMV GP-specific B cells become both activated and virally infected, making them targets for CTL killing. NO may have repressive influences on B cell activation or survival, whereas IL-10, TNF, and other cytokines may prompt the depletion of activated B cells through differentiation into apoptotic ASCs. The involvement of each repressive mechanism may differ depending on factors such as the route of infection, but the common outcome is an IFN-I–induced delay in the induction of long-lived neutralizing Ab responses.

The paper by Moseman et al. (3) also started with the discovery that IFN-I is a major factor repressing the early KL25HL B cell response. They depleted several cell types from LCMV-infected animals and, in their case, observed the most prominent rescuing effect after CD8 T cell depletion. In accord with CD8 T cells killing B cells, many GP-specific B cells are LCMV-infected. Remarkably, the CD8 T cell–mediated depletion of KL25HL B cells occurred as early as day 3, before considerable CTL expansion could have occurred. Infected B cells were protected in perforin-deficient mice, presumably due to compromised CTL function. In line with other studies, IFN-I was shown to promote the lytic function of CTLs. The authors next turned to imaging to determine whether CTL-B cell encounters could be visualized. They generated a new transgenic mouse line expressing a calcium sensor protein (GCaMP6s) and intercrossed this line with LCMV-specific TCR transgenic mice. Imaging of spleen slices from LCMV-infected mice that had received these T cells and fluorescent KL25HL B cells revealed CD8 T cells undergoing calcium fluxes after contact with the B cells. In transfers where the B cells expressed the calcium sensor, they were able to visualize cases of KL25HL B cells undergoing calcium flux due to CTL-induced loss of membrane integrity. Previous work implicated T cells in the elimination of LCMV neutralization-specific B cells, although this effect was later found to be variable and thus of unclear significance (1, 6). This study provides fresh evidence that CTL activity restrains the early response of B cells specific for LCMV GP and establishes an important role for IFN-I in augmenting this activity (Fig. 1).

Sammicheli et al. (4) began from a different angle by using intravital imaging to see whether they could visualize how LCMV thwarts the antibody response. To facilitate imaging, they used footpad inoculation rather than the more conventional intravenous infection route. They observed that LCMV replicates in subcaspular sinus and medullary macrophages in the draining lymph nodes before spreading by 72 hours to stromal cells in interfollicular areas. By contrast, VSV replication was restricted to subcapsular sinus macrophages and peaked by 8 to 12 hours. Tracking antigen-specific B cell behavior over 72 hours, VSV-specific cells behaved in accord with findings for nonreplicating antigens. However, the migration behavior of transferred LCMV-specific B cells was quite different. The cells showed a delayed movement out of the follicles, perhaps because of the slower kinetics of LCMV antigen accumulation, but they then showed a predilection for accumulation in interfollicular areas rather than along the B-T zone interface. Further tracking of response dynamics led the authors to find that LCMV infection caused a prominent and persistent recruitment of inflammatory monocytes. The LCMV-specific B cells underwent prolonged contacts with the inflammatory monocytes in interfollicular areas. IFN-I is an inducer of ligands for the monocyte chemokine receptor CCR2, and IFNAR-deficient mice showed reduced inflammatory monocyte recruitment and an augmented B cell response. Inflammatory monocytes are a source of nitric oxide (NO), and mice lacking inducible NO synthase (iNOS) in monocytes supported enhanced KL25HL B cell responses. A key conclusion from this study is that IFN-I–induced recruitment of inflammatory monocytes can hinder antiviral antibody responses in lymph nodes (Fig. 1).

In addition to establishing IFN-I as a repressor of early neutralizing antibody responses, the three studies are in accord in showing that any one effector does not account for the entire suppressive influence, making it clear that several factors cooperate. One limitation of the studies is their use of adoptive transfer of large numbers of high-affinity B cells, although key aspects of the findings are shown to hold up for endogenous B cell responses. Aside from the consistencies, there are some interesting discrepancies between the studies. Fallet et al. describe a mechanism of cell loss by differentiation to short-lived ASC, whereas the data of Moseman et al. suggest that many GP-specific B cells are directly killed. The latter effect is shown most clearly after the transfer of large numbers of LCMV-specific CD8 T cells. CD8 T cells can be a source of IL-10 and TNF, and it seems possible that the CD8 T cells contribute to B cell depletion through both direct killing and loss as apoptotic ASC. The requirements for B cell response suppression in the Sammicheli et al. study are the most distinct, with effects of CCR2 and iNOS deficiency and roles for inflammatory monocytes that are not seen in the other studies. These discrepancies most likely reflect the different routes of infection but might also be due to different doses of virus or slight differences in the viral GP. Differences in gut commensals might also be a factor because the microbiome can tune the systemic IFN-I pathway (7). Host genetics can have a major impact on how IFN-I acts during viral infections (8), something that should be kept in mind as this work is extended beyond the workhorse C57BL/6 mouse strain.

These important studies add to the growing body of evidence that despite having a critical protective influence early during viral infection, IFN-I actions can be counterproductive for the longer-term control of persistent viruses (5). It seems reasonable to speculate that viruses have taken advantage of this double-edged property of IFN-I to allow their escape from adaptive immunity. Of course, with new insight comes many new questions, including understanding how the suppressive effects on B cells are overcome such that a neutralizing antibody response is ultimately induced. This might reflect dissipation in IFN-I production and recruitment of newly generated B cells into the GC response. Given that chronic viral infections in humans are often associated with an IFN-I gene expression signature, it will be important to determine whether similar suppressive mechanisms are at play. Such understanding might then allow development of therapeutic approaches to augment neutralizing antibody responses while leaving the antiviral activities of IFN-I intact.


View Abstract

Stay Connected to Science Immunology

Navigate This Article