Requirement for the Transcription Factor LSIRF/IRF4 for Mature B and T Lymphocyte Function

See allHide authors and affiliations

Science  24 Jan 1997:
Vol. 275, Issue 5299, pp. 540-543
DOI: 10.1126/science.275.5299.540


Lymphocyte-specific interferon regulatory factor (LSIRF) (now called IRF4) is a transcription factor expressed only in lymphocytes. Mice deficient in IRF4 showed normal distribution of B and T lymphocyes at 4 to 5 weeks of age but developed progressive generalized lymphadenopathy. IRF4-deficient mice exhibited a profound reduction in serum immunoglobulin concentrations and did not mount detectable antibody responses. T lymphocyte function was also impaired in vivo; these mice could not generate cytotoxic or antitumor responses. Thus, IRF4 is essential for the function and homeostasis of both mature B and mature T lymphocytes.

Lymphocyte-specific interferon regulatory factor (LSIRF) [now called IRF4 (1)] is a lymphocyte-restricted member of the interferon regulatory factor (IRF) family of transcription factors (2, 3, 4). This family is defined by a characteristic DNA binding domain and the ability to bind to the interferon-stimulated response element. Members of the IRF family are involved in diverse processes such as pathogen response, cytokine signaling, apoptosis, and control of cell proliferation (5).

We generated mice deficient in IRF4 by replacing exons 2 and 3 of the IRF4 gene with a neomycin resistance gene (6). Mouse strains derived from two independent embryonic stem cell lines exhibited an identical phenotype. Mutation of the IRF4 gene was confirmed by Southern (DNA) blot analysis of tail DNA (shown for one mouse strain in Fig. 1A). Hind III-digested DNA from IRF4+/− and IRF4−/− mice displayed the 3.3-kb band of the mutant locus; the wild-type band at 8.4 kb was absent in IRF4−/− mice. The absence of the IRF4 protein was confirmed by protein immunoblot analysis (Fig. 1B).

Fig. 1.

Gene targeting of the murine IRF4 locus. (A) Southern blot analysis of mouse tail DNA. DNA was digested with Hind III and hybridized with the 5′ probe or with a neo probe. (B) Protein immunoblot analysis of concanavalin A (Con A)-activated lymph node cells with an antiserum specific for the COOH-terminus of IRF4 (2). Shown are nuclear extracts of IRF4+/+, IRF4+/−, and IRF4−/− lymph node cells. The IRF4 protein has a size of 51 kD and is indicated by the arrow. Cytoplasmic extracts were negative for IRF4 and are not shown. (C) Cellularity of spleen and lymph nodes in IRF4−/− mice. Cells from the thymus, spleen, and mesenteric lymph nodes were analyzed as described (16). IRF4+/+, triangles; IRF4+/−, diamonds; and IRF4−/−, solid circles.

At 4 to 5 weeks of age, lymph nodes and spleens of IRF4−/− mice showed a relatively normal lymphocyte distribution and cellularity as compared with those of control littermates (Fig. 1C). At 10 to 15 weeks, spleens were enlarged 3 to 5 times and lymph nodes were enlarged 10 times over those of control littermates, because of an expansion of T (both CD4+ and CD8+) and B lymphocytes (Fig. 1C). The distribution of several different Vβ-elements of the TCR was conserved, excluding the expansion of single T cell clones. Analysis of T cell surface molecules, including CD2, CD11a, CD18, CD25, CD28, CD45, CD54, FAS, and Thy-1, did not reveal any changes, although a slight increase in the number of CD69+ T cells was observed. Thymi of IRF4−/− mice were of normal size and showed a normal distribution of thymic cell populations (Fig. 1C) (7).

Analysis of B lymphocytes from bone marrow revealed no differences in the expression of the B cell surface molecules CD43, immunoglobulin M (IgM), IgD, Igκ, B220, and I-A, indicating that early B cell development was grossly normal. The development of peritoneal CD5+ B1 B cells was also normal (7). Splenic B cells showed normal surface expression of IgM and of κ and λ light chains (Fig. 2). However, on closer examination, spleens from IRF4−/− mice were found to display increased membrane IgM (mIgM)high mIgDlow, and decreased mIgMlow mIgDhigh, B cell populations. The frequency of CD23+ B220+ B cells was markedly reduced, and the CD23high B220+ B cell subpopulation was absent (Fig. 2), indicating a block at a late stage of peripheral B cell maturation (8). Consistent with such a block was the absence of germinal centers in B cell follicles of spleens and lymph nodes, even after the injection of sheep red blood cells, a stimulus that induces a large number of germinal centers in control mice (9). Furthermore, plasma cells could not be detected in the spleen or lamina propria of IRF4−/− mice.

Fig. 2.

Surface Ig and maturation markers in IRF4−/− B cells. Splenocytes were stained with fluorescein isothiocyanate—or phycoerythrin-conjugated mAbs to Igk, B220, CD23, IgD, and IgM or Abs to Igλ (anti-Igλ, Southern Biotech; all other mAbs, Pharmingen). Numbers indicate the percentage of cells in a quadrant or square. Staining and analysis were done as described (16).

Serum concentrations in IRF4−/− mice of all Ig subclasses were reduced at least 99% (Fig. 3A). Although control mice immunized with dinitrophenyl keywhole limpet hemocyanin (DNP-KLH) showed strong DNP-specific Ig production (Fig. 3B), IRF4−/− mice produced no detectable hapten-specific antibodies (Abs). Antibody production was also absent when IRF4−/− mice were immunized with the T-independent antigen trinitrophenyl-lipopolysaccharide (TNP-LPS) (Fig. 3C).

Fig. 3.

Basal serum Ig concentrations, and responses to T-dependent and T-independent antigens in IRF4−/− mice. (A) Serum concentrations of different Ig subclasses were determined in 6- to 8-week-old mice by enzyme-linked immunosorbent assay (ELISA) (16). IRF4+/+, triangles; IRF4+/−, diamonds; IRF4−/−, solid circles; ND, not detectable. (B) T-dependent antigens. IRF4+/− and IRF4−/− mice were injected subcutaneously with 100 μg of DNP-KLH in complete Freund's adjuvant and rechallenged on day 34 with 100 μg of DNP-KLH in incomplete Freund's adjuvant. Mice were bled, and hapten-specific IgG was determined by ELISA against DNP-bovine serum albumin (DNP-BSA). The results for day 44 are shown. A405, absorbance at 405 nm. (C) T-independent antigens. Mice were injected ip with 100 μg of TNP-LPS. After 5 days, TNP-specific IgM was determined by ELISA against TNP-BSA (17).

To examine B cell activation, we stimulated purified B cells from control and IRF4−/− mice with Ab to IgM (anti-IgM), anti-CD40, or LPS (Fig. 4A). Whereas control B cells showed a proliferative response after anti-IgM or LPS stimulation, proliferation of IRF4−/− B cells was reduced after LPS stimulation and absent after IgM stimulation. The addition of interleukin-4 (IL-4) could not restore the defective proliferation of IRF4−/− B cells. In contrast, stimulation with anti-CD40 induced an equal proliferative response in both IRF4+/+ and IRF4−/− B cells. Immunoglobulin secretion in response to stimulation with LPS or monoclonal antibody (mAb) to CD40 was severely impaired in IRF4−/− B cells (Fig. 4B).

Fig. 4.

(A) Proliferation of purified B cells after stimulation with IL-4, anti-IgM, LPS, or anti-CD40. Splenic B cells were purified by erythrocyte lysis and complement lysis of T cells as described (18). B cells (5 × 105 per well) were stimulated with LPS (10 μg/ml), goat anti-mouse IgM (20 μg/ml), anti-CD40 (5 μg/ml, clone 3/230), or mouse recombinant IL-4 (50 U/ml). Proliferation was measured at day 3 by [3H]thymidine incorporation (16). Values are the mean of triplicate samples. (B) Antibody production after B cell stimulation with LPS and mAb to CD40 in vitro. Conditions were as described in (A). Total Ig was measured by ELISA after 3 days (16). Each experiment in (A) and (B) is representative of at least three independent experiments.

IRF4 is induced in T cells after T cell receptor stimulation (2). To examine the in vitro response of IRF4−/− T cells, we incubated lymph node cells with mAb to CD3, concanavalin A, or the bacterial superantigen staphylococcal enterotoxin A (SEA) and measured cell proliferation (Fig. 5A). Proliferation of IRF4−/− T cells was reduced after incubation with all three stimuli and was not restored by the addition of IL-2. In mixed lymphocyte reaction (MLR) experiments, purified IRF4−/− T cells (H-2b) showed reduced proliferation against BALB/c spleen cells (H-2d) (Fig. 5B), which is consistent with a defect intrinsic to IRF4−/− T cells. Furthermore, IRF4−/− spleen and peritoneal cells were able to process and present antigens to wild-type T cells (10), excluding impaired antigen presentation of IRF4−/− cells. When IRF4−/− T cells were stimulated with mAb to CD3, the production of IL-2, IL-4, and interferon-γ (IFN-γ) was also found to be reduced (Fig. 5C).

Fig. 5.

In vitro and in vivo T cell responses of IRF4−/− mice. (A) Proliferation of lymph node T cells after stimulation with mAb to CD3ε, Con A, and SEA. (B) Mixed lymphocyte reaction. (C) Cytokine production after stimulation with mAb to CD3ε. (D) Cytotoxic response against LCMV (the control was an uninjected mouse). Proliferation after stimulation with Con A (2 μg/ml), anti-CD3ε (clone 145 2C11, 10 μg/ml), SEA (1 μg/ml), and IL-2 (50 U/ml) and MLR assays were performed as described (16, 19). For the MLR assays, T cells were purified with T cell enrichment columns (R&D Systems). For the induction of lymphokines, lymph node cells (4 × 106 per milliliter) were cultured for 48 hours in plates coated with anti-CD3ε (5 μg/ml) (15). IL-2 was determined with the CTLL-2 bioassay (16) and IL-4 and IFN-γ by ELISA (Genzyme). Values are the mean of triplicate samples. For determination of cytotoxic responses, mice were footpad-injected with 200 plaque-forming units of LCMV (Armstrong strain). Ex vivo cytotoxic activity of T cells was measured 8 days after infection in a 51Cr-release assay with EL-4 target cells loaded with LCMV-glycoprotein peptide (amino acids 33 to 41) (19). Experiments in (A) through (D) are representative of at least three independent experiments, with at least three mice per group in (D).

To analyze the function of T cells in vivo, we infected IRF4−/− mice with lymphocytic choriomeningitis virus (LCMV). When injected into the footpads of normal mice, LCMV induces a strong cytotoxic T cell response, characterized by an immunopathologic swelling reaction which, in the early phase, is mediated by CD8+ T cells (11). LCMV was injected into the footpads of mice, and after 8 days, spleen cells were analyzed for LCMV-specific cytotoxic activity (Fig. 5D). In contrast to cells from control mice, cells from IRF4−/− mice had no cytotoxic activity. In addition, whereas control mice showed the expected strong swelling of their footpads, swelling was absent in IRF4−/− mice.

The transplantation of T cells into an immunocompromized allogeneic host leads to a strong T cell-mediated graft-versus-host (GvH) reaction. BALB/c mice (H-2d) with severe combined immunodeficiency disease (SCID) injected with spleen cells from allogeneic IRF4+/+ and IRF4+/− mice (all H-2b) developed a severe GvH reaction, accompanied by rough fur, overwhelming diarrhea, and massive weight loss (12). These mice died within 18 days of injection. Although SCID mice injected with allogeneic IRF4−/− cells displayed an initial weight loss, they recovered and survived for at least 40 days. Flow cytometric analysis revealed that after 4 weeks, the peripheral blood of these mice contained H-2Kb+ donor-derived lymphocytes.

Finally, mice were injected with P815 mastocytoma cells, derived from DBA/2 mice (H-2d). When injected into H-2b mice, P815 cells are completely allogeneic and provoke a strong T cell response. All IRF4+/+ and IRF4+/− mice were able to reject the P815 cells, and no tumor growth occurred for at least 3 months. In contrast, all IRF4−/− mice developed tumors with a time course identical to that seen in DBA/2 control mice (13).

It has been shown that IRF4 binds a motif in the Ig light chain gene enhancers Eκ3′, Eλ2−4, and Eλ3−1 and that IRF4 can enhance the transcription of genes under the control of this motif (3). For the κ-gene locus it has further been demonstrated that Eκ3′ is crucial for κ-light chain secretion (14). Together, these facts could explain the defective Ig production in IRF4−/− mice. However, it is not likely that impaired light chain expression alone accounts for the severe B cell defect, and it is probable that IRF4 is involved in the expression of other genes that are important for late B cell differentiation.

In T cells, the deficiency of IRF4 leads to reduced proliferation and lymphokine production, and the reduced proliferation cannot be restored by exogenous IL-2. Analysis of early events after T cell activation, such as calcium influx or the expression of the activation molecules CD25 and CD69, revealed no difference between IRF4−/− and control T cells (7). Therefore, we conclude that IRF4 is not involved in the early events after T cell activation but is important for later processes, including both IL-2 production and IL-2 response, which is consistent with the strong up-regulation of IRF4 mRNA shortly after T cell activation (2).

Paradoxically, even though B and T cell responses are impaired in IRF4−/− mice, these animals develop severe lymphadenopathy. A possible explanation is that the incomplete lymphocyte activation in IRF4−/− mice affects the mechanisms controlling lymphocyte homeostasis, a phenomenon that has been described for IL-2- and IL-2Rα-deficient mice (15).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
View Abstract

Navigate This Article