Virus-induced transient bone marrow aplasia: major role of interferon-alpha/beta during acute infection with the noncytopathic lymphocytic choriomeningitis virus

Abstract

The hematologic consequences of infection with the noncytopathic lymphocytic choriomeningitis virus (LCMV) were studied in wild-type mice with inherent variations in their interferon (IFN)-alpha/beta responder ability and in mutant mice lacking alpha/beta (IFN-alpha/beta R0/0) or gamma IFN (IFN-gamma R0/0) receptors. During the first week of infection, wild type mice demonstrated a transient pancytopenia. Within a given genetic background, the extent of the blood cell abnormalities did not correlate with the virulence of the LCMV isolate but variations were detected between different mouse strains: they were found to depend on their IFN-alpha/beta responder phenotype. Whereas IFN-gamma R0/0 mice were comparable to wild-type mice, IFN-alpha/beta R0/0 mice exhibited unchanged peripheral blood values during acute LCMV infection. In parallel, the bone marrow (BM) cellularity, the pluripotential and committed progenitor compartments were up to 30-fold reduced in wild type and IFN-gamma R0/0, but remained unchanged in IFN-alpha/beta R0/0 mice. Viral titers in BM 3 d after LCMV infection were similar in these mice, but antigen localization was different. Viral antigen was predominantly confined to stromal BM in normal mice and IFN-gamma R0/0 knockouts, whereas, in IFN-alpha/beta R0/0 mice, LCMV was detected in > 90% of megakaryocytes and 10-15% of myeloid precursors, but not in erythroblasts Although IFN-alpha/beta efficiently prevented viral replication in potentially susceptible hematopoietic cells, even in overwhelming LCMV infection, unlimited virus multiplication in platelet and myeloid precursors in IFN-alpha/beta R0/0 mice did not interfere with the number of circulating blood cells. Natural killer (NK) cell expansion and activity in the BM was comparable on day 3 after infection in mutant and control mice. Adaptive immune responses did not play a major role because comparable kinetics of LCMV-induced pancytopenia and transient depletion of the pluripotential and committed progenitor compartments were observed in CD8(0/0) and CD4(0/0) mice, in mice depleted of NK cells, in lpr mice, and in perforin-deficient (P0/0) mice lacking lytic NK cells. Thus, the reversible depression of hematopoiesis during early LCMV infection was not mediated by LCMV-WE-specific cytotoxic T lymphocyte, cytolysis, or secreted IFN-gamma from virally induced NK cells but was a direct effect of IFN-alpha/beta.

Figure 1

Kinetics of BM cellularity after LCMV-WE infection. Mice of wild type genotype (open columns) or mutant mice with inactivated α/β IFN (closed columns) or γ IFN (hatched columns) receptors were infected intravenously with LCMV-WE (2 × 10⁶ PFU). At the indicated timepoints, mice were killed and the absolute number of nucleated BM cells per femur was determined. The data represent mean ± SD of 3–4 mice per group.

Figure 2

Lineage-committed precursor cells in wild-type (open columns) or mutant (closed columns, IFN-α/β R^0/0; hatched columns, IFN-γ R^0/0) mice 3 d after LCMV-WE (2 × 10⁶ PFU) infection. Results are presented as the mean number (± SD) of BFU-E, CFU-GM, or CFUMeg for duplicate methylcellulose cultures of total BM cells per femur. Pooled data from two independent experiments with 2–3 individual mice per group are shown.

Figure 3

Number of pluripotential hematopoietic progenitors in the BM 3 d after infection with LCMV-WE (2 × 10⁶ PFU). 10⁵ syngeneic BM cells of uninfected (open symbols) or infected (closed symbols) wild-type (129Sv/Ev, H-2^b) or mutant (IFN-α/β R^0/0 or IFN-γ R^0/0, H-2^b) donor mice were injected intravenously into lethally irradiated LCMV-immune recipient wild-type mice (129Sv/Ev, H-2^b). CFU-S in the spleens were determined 8 d later. Each dot shows the number of colonies of an individual recipient. The horizontal lines represent the mean number of colonies per femur transferred from one individual donor mouse. The mean CFU-S of two individual donor mice per group are shown.

Figure 4

Time kinetics of virus titers in the BM after infection with LCMV. Mice were infected intravenously with LCMV-WE (2 × 10⁶ PFU) and virus titers were determined in 10⁶ nucleated BM cells at the timepoints indicated. Data points are values for individual mice and the horizontal lines represent the mean of the indicated group.

Figure 5

(A) Infection of stromal cells in wild-type and IFN-γ R^0/0 mice versus blood precursor cells in IFN-α/β R^0/0 mice. Viral protein in differentiated BM cells was detected by immunocytochemical staining 1 d (A–C) and 3 d (D–F) after infection with 2 × 10⁶ PFU LCMV-WE. BM smears were stained with a LCMV-specific mAb followed by a peroxidase-labeled goat anti–rat Ig and by an alkaline phosphatase–labeled donkey anti–goat Ig. LCMV-WE–infected cells are colored red and uninfected cells appear blue due to counterstaining with Meyer's hemalum. Note the intense staining of viral antigen in megakaryoctes and, to a lesser extent, in myeloid precursors and stromal cells in IFN-α/β R^0/0 mice (B, E). In wild-type and IFN-γ R^0/0 mice, LCMV-WE is detected in substantial amounts exclusively in stromal fibroblast and endothelial cells (A, D, C, F). Original magnification, (A–C) ×500; (D–F) ×1,000. (B) Differential quantification of LCMV-infected BM cells. Fixed BM smears of mice of the indicated genotypes were stained with a LCMV-specific mAb at different timepoints after infection with LCMV-WE. The frequency of infected cells was determined microscopically by counting 200 cells per individual mouse. Bars represent the mean ± SD of the percentage of WE-positive BM cells of 2–3 individual mice per group.

Figure 5

(A) Infection of stromal cells in wild-type and IFN-γ R^0/0 mice versus blood precursor cells in IFN-α/β R^0/0 mice. Viral protein in differentiated BM cells was detected by immunocytochemical staining 1 d (A–C) and 3 d (D–F) after infection with 2 × 10⁶ PFU LCMV-WE. BM smears were stained with a LCMV-specific mAb followed by a peroxidase-labeled goat anti–rat Ig and by an alkaline phosphatase–labeled donkey anti–goat Ig. LCMV-WE–infected cells are colored red and uninfected cells appear blue due to counterstaining with Meyer's hemalum. Note the intense staining of viral antigen in megakaryoctes and, to a lesser extent, in myeloid precursors and stromal cells in IFN-α/β R^0/0 mice (B, E). In wild-type and IFN-γ R^0/0 mice, LCMV-WE is detected in substantial amounts exclusively in stromal fibroblast and endothelial cells (A, D, C, F). Original magnification, (A–C) ×500; (D–F) ×1,000. (B) Differential quantification of LCMV-infected BM cells. Fixed BM smears of mice of the indicated genotypes were stained with a LCMV-specific mAb at different timepoints after infection with LCMV-WE. The frequency of infected cells was determined microscopically by counting 200 cells per individual mouse. Bars represent the mean ± SD of the percentage of WE-positive BM cells of 2–3 individual mice per group.

Figure 6

NK activity in the BM of acutely infected (day 3) IFN R-deficient/competent mice. Effector cells of LCMV-WE–infected (closed triangle) or uninfected (open triangle) or poly I/C-treated (closed circle) mice were incubated for 5 h with ⁵¹Cr-labeled NK-sensitive YAC-1 cells. LUs were determined as the number of lymphocytes necessary to lyse 33% of the target cells during the standard test duration. The indicated number represents the LU per BM of the lower extremities and was computed by dividing the total number of BM cells by LU. Spontaneous ⁵¹Cr release was <25%.

Figure 7

Expansion of NK cells in the BM in response to LCMV studied by FACS^® analysis 3 d after infection with LCMV-WE (2 × 10⁶ PFU). NK cells in the BM were detected by positive expression of the NK 1.1 marker and absence of CD8 and CD4. CD4/CD8 (FITC) and NK 1.1 (PE) staining is shown for cells electronically gated on the erythroid (TER 119 Tricolor) negative BM population. Percentages are given for dot plot quadrants. Plots are repesentative for three mice per group.