Type 1 IFN Deficiency in the Absence of Normal Splenic Architecture during Lymphocytic Choriomeningitis Virus Infection1

The innate immune system uses different mechanisms to respond to infectious pathogens. Experiments evaluating the requirements for a type 1 IFN (IFN-αβ) response to lymphocytic choriomeningitis virus (LCMV) resulted in the surprising discovery that mice deficient in B and T cell development, i.e., RAG-deficient and SCID, had profoundly reduced levels of IFN-αβ in serum and spleen, despite high viral replication. In addition to lacking an adaptive immune system, these strains exhibit aberrant splenic architecture, and the defect in type 1 IFN production was also observed in mice lacking normal splenic marginal zone (MZ) organization due to genetic deficiencies in B cell development or in cytokine functions required for development of the MZ, i.e., μMT, lymphotoxin-α, and TNFR1. Interestingly, the IFN-αβ reduction was not observed after murine CMV infection. Depletion of phagocytic cells from normally developed spleens by treatment with clodronate-containing liposomes demonstrated that these populations were required for the type 1 IFN response to LCMV, but not to murine CMV, and for control of viral replication. Complete repopulation of the MZ was necessary to restore normal IFN-αβ production. In contrast, control of LCMV replication correlated with the return of CD11c+ cells. Taken together, these results demonstrate the complexity and sophistication of the splenic MZ in sensing and responding to particular pathogens and reveal the importance of organ architecture in the production of type 1 IFN.

T ype 1 IFNs, IFN-␣␤, are inhibitors of viral infection but also have immunoregulatory functions, including induction of dendritic cell (DC) 3 maturation and activation (1)(2)(3) and promotion of CD8 T cell responses (4 -7). These innate cytokines can be produced by a wide range of cell types and are induced to high levels in the spleens of mice infected with a variety of viruses including lymphocytic choriomeningitis virus (LCMV) and murine CMV (MCMV) (8,9). The spleen is important for defense against systemic infections, and the specialized organization of its marginal zone (MZ) facilitates filtering of blood-borne pathogens, as well as processing and presentation of Ag for initiation of adaptive immune responses (10). The MZ, situated between red and white pulp regions, is histologically defined by sinus-lining mucosal addressin cell adhesion molecule-1 ϩ (MAdCAM-1 ϩ ) reticular endothelial cells, metallophilic macrophages (MM) that stain with the MOMA-1 Ab, and specific intracellular adhesion molecule-grabbing nonintegrin receptor 1 ϩ (SIGN-R1 ϩ ) MZ macrophages (MZM). Although many cell types traffic through the area, MZ B cells and CD11b ϩ CD11c ϩ myeloid DCs are interspersed within the region (10 -12). Many viruses initiate early infection and replication within MZ cells, and type 1 IFN expression can be demonstrated in the MZ under these conditions (8,9).
The studies presented here were undertaken to further define the requirements for type 1 IFN production during LCMV infection. Surprisingly, mice lacking functional splenic architecture, as a result of genetic deficiencies in B cells, lymphotoxin ␣ (LT␣), or TNF signaling, exhibited profound reductions in type 1 IFN during LCMV infection. In contrast, responses against MCMV were normal. Treatments with clodronate-containing liposomes to deplete phagocytic cells from a normally developed spleen demonstrated that these populations were required for type 1 IFN production in response to LCMV but not MCMV. Repopulation experiments showed that restoration of the CD11c ϩ , MOMA-1 ϩ , and F4/80 ϩ populations was not sufficient to reconstitute the type 1 IFN response to LCMV, but that complete repopulation, with restoration of SIGN-R1 ϩ cells, was required. In contrast, control of LCMV replication correlated with the return of CD11c ϩ cells to the MZ. Taken together, these results further demonstrate the complexity and sophistication of the splenic MZ for sensing and responding to a variety of pathogens. The work has important implications for the involvement of splenic architecture in the production of type 1 IFN, and the consequences of deficiencies on the subsequent activation of vigorous DC and CD8 T cell responses during particular viral infections.

In vivo treatments and sample collection
Viral infections were initiated at day 0 by i.p. injection of 2 ϫ 10 4 PFU of LCMV-Armstrong clone E350, 2 ϫ 10 4 PFU of LCMV-WE, 10 4 PFU of MCMV Smith (V70) strain, or by i.v. injection of 2 ϫ 10 6 PFU LCMVclone 13. Phagocytic cells of the spleen were depleted by treating mice with liposomes containing dichloromethylene bisphosphonate (clodronate liposomes), prepared as previously described (14,15). Clodronate was a gift from Roche. Mice received one injection of 0.2 ml of clodronate liposomes i.v. at day Ϫ1 before infection. Where indicated, mice received clodronate liposome injection at day Ϫ1 and were infected on the specified day. As expected, clodronate liposome treatment depleted MOMA-1 ϩ MMs, SIGN-R1 ϩ MZMs, F4/80 ϩ red pulp macrophages, and MZ CD11c ϩ cells, as assessed by immunohistochemical analyses (15,16). Flow cytometric analyses of isolated splenic populations following treatment indicated that plasmacytoid DCs (PDCs) were reduced by Ͻ25%. PBS liposomes were injected in the same manner into control mice, with similar results as unmanipulated mice (data not shown). Sera and spleen homogenates were prepared as described (9).

Quantitation of type 1 IFN levels and viral titers
Type 1 IFN bioactivity in sera and spleen homogenates was detected by a biological assay against vesicular stomatitis virus (VSV) (8,17). VSV was titrated against a type 1 IFN reference standard to determine the dose at which 50% protection from lysis corresponded to 1 U/ml. The limit of detection for serum and spleen bioassays was 8 U/ml and 8 U/(spleen weight), respectively. IFN-␣ was determined by ELISA, as described previously (8); the limit of detection for serum and spleen was 780 pg/ml and 780 pg/(spleen weight), respectively. LCMV or MCMV viral titers were determined by plaque assays on Vero cells or mouse embryo fibroblasts (6,8,18). The limit of detection was 100 PFU and 100 PFU/(spleen weight) for serum and spleen viral titers.

Immunohistochemical staining
Spleens were embedded in OCT (Sekura Finetek), cut into 5-m sections with a cryostat (CM3050S; Leica), fixed in acetone, and rehydrated. Subsequently, sections were stained for MAdCAM-1, CD11c, F4/80 (all using Abs from eBioscience), SIGN-R1 (Novus Biologicals), or MOMA-1 (Serotec). Biotinylated secondary reagents were purchased from Jackson Im-munoResearch Laboratories. Alkaline phosphatase Vectastain ABC reagents, Vector Blue substrate, and Nuclear Fast Red counterstain were purchased from Vector Laboratories. Magnifications are ϫ10 for all sections with the exception of inset magnifications, which are ϫ27.5. Photographs were taken with a Cool Snap camera and RSImage software (Roeper Scientific).

Effects of T and B cell deficiencies on the type 1 IFN response to LCMV
Genetically mutated mice lacking various cell populations were examined to help characterize the requirements for type 1 IFN responses to LCMV. For these studies, mice were challenged with LCMV-Armstrong (herein referred to as LCMV), as well as the more virulent strains LCMV-WE and -clone 13 (19). Serum samples and spleen homogenates were prepared to quantitate the in vivo production of type 1 IFN. A biological assay was used to measure the total IFN-␣␤ response based on units, and an IFN-␣specific ELISA was used to quantify the production of this cytokine in nanograms. Surprisingly, C57BL/6 mice that lack T and B cells as a result of a deficiency in Rag1 (RAG-1 Ϫ/Ϫ ) produced low to undetectable levels of type 1 IFN in spleen homogenates ( Fig.  1A) after infection with all three LCMV isolates. In comparison to control C57BL/6 mice that produced a robust type 1 IFN response sustained for up to 4 days, the RAG-1-deficient mice had 25-to 200-fold lower responses at the peak of production. SCID mice, also lacking T and B cell populations, displayed similar deficiencies (Fig. 1B). LCMV replication was slightly delayed in both strains of mice ( Fig. 1, C and D), but a normal type 1 IFN response never occurred, even when levels of viral replication surpassed those detected during infections of control mice. Similar cytokine results were obtained with serum samples (Fig. 1, E and F). Thus, the absence of T and B cells renders the host profoundly impaired in its ability to mount a type 1 IFN response to LCMV.

Splenic architecture requirements for the type 1 IFN response to LCMV
To determine which absent cell population was responsible for the reduction in type 1 IFN, mice lacking either the T or B cell population individually were examined. TCR-␤ Ϫ/Ϫ and nude mice, specifically lacking T cells, or E26 mice, lacking T and NK cells, displayed normal amounts of type 1 IFN in serum or spleen homogenates at day 2 of LCMV infection and showed normal levels of LCMV replication at this time during infection (Table I). Thus, T and NK cells are not required for type 1 IFN production during LCMV infection. However, MT mice, specifically lacking B cells as a result of IgM H chain gene mutation, produced undetectable levels of type 1 IFN despite clear viral replication (Table  I). To determine whether B cells were the major IFN-␣␤ producers under normal conditions, splenic B cells were isolated from LCMV-infected control mice and evaluated for their ability to produce the cytokine ex vivo. The enriched B cell population did not produce increased levels of type 1 IFN in 24-h conditioned medium (data not shown). Taken together, these data show that B cell-deficient mice are not able to initiate a type 1 IFN response, but B cells themselves do not appear to be the direct producers of the cytokine.
As B cells are involved in the expression of LT␣ 1 ␤ 2 and TNF, and as these cytokines are important for the development of a spleen with normal organization, B cell-deficient mice exhibit defective splenic architecture and organogenesis (10,20). To evaluate the consequences of disrupted splenic organization in the presence of B cell populations, LT␣-or TNFR1-deficient mouse strains, known to have disorganized white pulp compartments and poorly developed MZs, were examined. After infection with LCMV for 2 days, these mutated mice showed an inability to produce normal amounts of type 1 IFN despite increasing viral titers in the spleen (Table II). These data demonstrate that normal splenic architecture is critical for the type 1 IFN response to LCMV.

Different type 1 IFN responses to LCMV or MCMV in mice with aberrant splenic architecture
To examine the consequences of disrupted splenic organization on the host response to other viruses, responses to MCMV were contrasted to LCMV. As compared with mice on C57BL/6 backgrounds, 129 mice mount much greater innate cytokine responses to MCMV, with peak levels achieved at 1.5 days after infection (8). In carrying out the experiments reported here, it became apparent that the converse is true for responses to LCMV:C57BL/6 mice produce much more type 1 IFN than 129 mice against LCMV. To contrast any differences in type 1 IFN production against the two viruses, we took advantage of mice deficient in Rag2 (RAG-2 Ϫ/Ϫ ) because they are available on both genetic backgrounds. Similarly to RAG-1 Ϫ/Ϫ or SCID mice, RAG-2 Ϫ/Ϫ mice do not generate mature B and T cells, and failed to generate a robust type 1 IFN response to LCMV when examined on a C57BL/6 background ( Fig. 2A). In contrast, RAG-2 Ϫ/Ϫ 129 mice infected with MCMV had normal or greater than normal type 1 IFN responses at peak times after challenge. Viral replication was within 1 log under all conditions tested. Similar cytokine results were obtained with serum samples (Fig. 2B). Responses of RAG-2 Ϫ/Ϫ mice on the 129 background against LCMV, or on the C57BL/6 background against MCMV, were similar to those presented in Fig. 2A, but maximal cytokine levels achieved were ϳ10-fold lower than in the high responder strains (data not 14 BLD a Studies with genetically mutated mice were done as described in Materials and Methods. Serum samples and spleen homogenates from day 2 LCMV-infected mice were assayed for IFN-␣␤ using a biological assay. Spleen homogenates and serum were assayed for LCMV by plaque assays on Vero cells. Results are expressed as means Ϯ SEM of three to four mice per group. Uninfected mice did not produce detectable type 1 IFN. BLD, below limits of detection. shown). These results indicate that the rigid requirement of splenic architecture in supporting type 1 IFN responses is specific to some, but not all, infectious agents. These data also highlight that type 1 IFNs are likely induced by different mechanisms based on the specific pathogen encountered.

Depletion and reconstitution of type 1 IFN responses following elimination of phagocytic cells
The effects of architecture could be a result of structural components and/or differences in cellular composition resulting from changes in structure. The highly organized MZ area is composed of many different cell types including sinus-lining MAdCAM-1 ϩ reticular endothelial cells, MOMA-1 ϩ MMs, SIGN-R1 ϩ MZMs, MZ B cells, PDCs, and CD11c ϩ CD11b ϩ myeloid DCs (10). To further delineate whether specific MZ constituents are responsible for the lack of type 1 IFN production during LCMV infection, clodronate-containing liposomes ("clodronate liposomes") were administered to control mice (14). Because clodronate liposomes are engulfed by phagocytic cells, this apoptotic agent can be used to deplete MZMs (characterized by SIGN-R1 ϩ staining), MMs (MOMA-1 ϩ ), red pulp macrophages (F4/80 ϩ ), and the phagocytic DCs located in the bridging channels of the MZ (CD11c ϩ ) from normally developed spleens, leaving the rest of the spleen architecture mostly unperturbed. Treatment with clodronate liposomes alone did not induce a type 1 IFN response in uninfected mice, and cytokine responses induced by LCMV infection were similar between unmanipulated mice and mice treated with control PBScontaining liposomes. In contrast, administration of clodronate liposomes to C57BL/6 mice 1 day before LCMV infection dramatically reduced the levels of type 1 IFN in spleen homogenates (Fig. 2C) at the day 2 peak of production, and at all later times tested during infection (0% of control at days 4 and 5). Clodronate liposome treatment had only modest effects at the peak of type 1 IFN production during MCMV infection of 129 mice, consistent with the slight reduction in PDCs and with the normal levels of IFN-␣␤ observed against MCMV in mice lacking proper splenic architecture. Administration of clodronate liposomes during either viral infection resulted in increased splenic viral titers (Fig. 2C). Serum samples gave similar type 1 IFN results (Fig.   2D). Thus, specific phagocytic MZ cells are imperative for the full type 1 IFN response to LCMV, but not to MCMV.
Short-term experiments using splenic leukocytes or bone marrow were unsuccessful at reconstituting the depleted MZ populations, so repopulation studies were undertaken. Because clodronate-depleted subsets can be endogenously repopulated with different kinetics (14), studies were conducted to determine at what time following their administration type 1 IFN production was restored. Groups of C57BL/6 mice were control-treated or injected with clodronate liposomes (at day Ϫ1), then left uninfected or LCMV infected 1 day later (at day 0) or at weekly intervals following treatment. Immunohistochemical analyses were performed to evaluate the frequencies of particular cell subsets, and type 1 IFN levels in the spleen were evaluated at day 2 following LCMV infection. As expected, uninfected control mice showed MZs composed of MAdCAM-1 ϩ , MOMA-1 ϩ , and SIGN-R1 ϩ cells, with F4/80 ϩ macrophages in red pulp, and CD11c ϩ cells in MZs and white pulp. During infection, MAdCAM-1 ϩ expression was largely unchanged, whereas SIGN-R1 ϩ staining was reduced and MOMA-1 ϩ staining extended into red pulp, F4/80 ϩ staining became more intense, and CD11c ϩ cells extended into the red pulp (Fig. 3B). Clodronate liposome treatment did not affect the MAdCAM-1 ϩ reticular endothelial cells, as expected, but depleted the majority of MM, MZM, and red pulp macrophages. The MZ CD11c ϩ cells were depleted, but those in the white pulp remained. LCMV infection of clodronate liposome-treated animals did not induce the reappearance of the depleted cell populations. Type 1 IFN production was still absent in LCMV-infected mice rested for 7, 14, 21, 28, and 35 days following clodronate liposome administration before infection (data not shown). By 42 days following treatment, all examined depleted cell populations had reappeared, with the exception of SIGN-R1 ϩ MZMs that were still at Ͻ25% of control values. Nevertheless, the type 1 IFN responses remained largely absent during LCMV infection (Fig. 3, B and A). However, when mice were infected at day 63 following liposome administration, IFN-␣␤ production had returned to control levels (Fig. 3A). This was associated with a Ͼ4-fold increase in SIGN-R1 ϩ staining to near control levels; the numbers of MZM cells per section in the control group at day 63 being 432 (Ϯ22) and in the reconstituted Control mice (f) or RAG-2 Ϫ/Ϫ mice (Ⅺ) were uninfected or infected with LCMV-Armstrong or MCMV for indicated times. C and D, Uninjected mice (f) or mice receiving an i.v. injection of clodronate liposomes 1 day before infection (Ⅺ) were infected with LCMV or MCMV at day 0 for indicated times. Levels of IFN-␣␤ or IFN-␣ viral titers were determined on spleen homogenates by bioassay, sandwich ELISA, or plaque assay, respectively (A and C). Levels of IFN-␣␤ in serum samples were determined by bioassay (B and D). Results are presented as means Ϯ SEM of four to seven mice per group and representative of two to four independent experiments. , Below limit of detection. group being 410 (Ϯ2) (Fig. 3B). Thus, either SIGN-R1 ϩ cells are producing type 1 IFN during LCMV infection, or the entire MZ architecture must be intact for normal IFN-␣␤ responses to LCMV.

Role of the MZ in control of LCMV replication and dissemination
A clear dichotomy was observed between mice lacking proper architecture and clodronate liposome-treated mice. Mice with aberrant architecture (SCID or TNFR1 Ϫ/Ϫ , for example) showed similar or slightly lower viral titers in the spleen at day 2 of LCMV infection, and no observable serum viremia (Tables I and II, and Fig. 4A). However, C57BL/6 mice receiving clodronate liposomes at 1 day before infection show more than a log increase in LCMV replication over control mice in the spleen and dissemination to the serum occurs, which is normally undetectable in control mice. Interestingly, clodronate liposome-treated mice allowed to rest 14 days (2 wk) before infection exhibit normalized splenic viral titers and spread to the serum is not observed (Fig. 4A). A time course showed that viremia was still observable and splenic viral titers remained higher than in control mice when LCMV infection was initiated at day 7, but not at day 14, following liposome treatment (Fig. 4B). Immunohistochemical studies were performed on sections from these mice to examine which depleted cell populations had returned to the spleens of mice infected at day 14 compared with those infected at day 7 following liposome administration. As shown in Fig. 4C, CD11c ϩ cells, and F4/80 ϩ cells, to a lesser extent, had repopulated the spleen by day 14. In contrast, MMs and MZMs had not returned. Thus, even though the depleted cell types that have repopulated at 14 days following liposome treatment are not able to support type 1 IFN production, they are sufficient to control local spleen viral replication and prevent dissemination to the serum. Taken together, these results indicate that early defense against infection can be controlled by part of the MZ constituents, but the entire MZ must be intact for a complete type 1 IFN response to LCMV.

Discussion
The data presented herein demonstrate that splenic architecture, and the MZ in particular, are important for defense against infection with viruses. They indicate that there are two different mechanisms influencing protection: promotion of type 1 IFN responses and control of viral replication. The surprising observation that MZ organization, and MZMs, specifically, are critically involved in the production of type 1 IFN against LCMV is reported here for the first time. After depletion of phagocytic cells by clodronate liposomes, type 1 IFN production was largely absent and only returned following infection after extended recovery, correlating with the restoration of SIGN-R1 ϩ MZMs. PDCs were mostly unaffected by the treatment and other CD11c ϩ DC subsets returned at earlier times after treatment. Thus, the results clearly show that although FIGURE 3. The splenic MZ must be intact for proper type 1 IFN responses to LCMV. A, C57BL/6 mice were uninjected (f) or i.v. injected with clodronate liposomes (Ⅺ) at day Ϫ1 and left for the indicated days before being infected i.p. with LCMV. All mice were harvested at day 2 following infection. -, Mice were left uninfected. Levels of IFN-␣␤ or IFN-␣ were determined on spleen homogenates by bioassay and sandwich ELISA, respectively. The percentage of type 1 IFN production in clodronate liposome-treated animals as compared with WT controls is noted. B, Immunohistochemical analysis of spleen sections for MAdCAM-1, CD11c, F4/ 80, MOMA-1, and SIGN-R1 from control or liposome-injected mice either uninfected or infected with LCMV for 2 days. "Infection Day," day of infection, where depletion occurred at day Ϫ1 and infection occurred at day listed (0, 42, or 63) (see Materials and Methods); "UN," mice were left uninfected; "Clodronate Liposomes," whether clodronate liposomes were administered. Magnification is ϫ10, except for the magnified presentation of the identified insets of SIGN-R1 staining. Results are presented as means Ϯ SEM of two to five mice per group and are representative of at least two independent experiments. , Below limit of detection.
DCs may be induced to produce type 1 IFN in response to LCMV (21,22), they are not sufficient to initiate the response in the absence of the macrophages of the MZ. A similar scenario occurred in mice known to lack proper splenic architecture, i.e., RAG Ϫ/Ϫ , SCID, LT␣ Ϫ/Ϫ , or TNFR1 Ϫ/Ϫ . Because the type 1 IFNs have potent antiviral and immunoregulatory functions, the data have significant implications for the understanding of defense against viral infections. They show that the structure and composition of the splenic MZ, critical in protecting against systemic infections and for processing and presentation of Ags to initiate adaptive immune responses, is also key for induction of these innate cytokines.
The results emphasize the importance of non-PDC populations for type 1 IFN induction in response to certain viral infections. The experiments presented here do not specifically define the cellular sources of the cytokines during LCMV infection, but are consistent with previous work from our group demonstrating that PDCs are the major producers during MCMV but not LCMV infection (8), as well as other reports implicating a role for non-PDCs in the expression of type 1 IFN in response to HSV infection in the mouse and to Sendai viral challenge of human leukocytes in culture (23)(24)(25). Interestingly, the PDC-dependent type 1 IFN response to MCMV infection did not have the same strict requirement for splenic architecture (Fig. 2). Thus, the sensing of these two infections is fundamentally different. In the case of LCMV infection, the complete MZ structure may either 1) supply the cells directly producing type 1 IFN, and/or 2) provide necessary support for intercellular communication for induction of the peak type 1 IFN response. Ongoing studies are dissecting these possibilities. Nevertheless, the results presented here make the novel and important discovery of the structural contribution to the innate cytokine responses against a virus.
Consistent with other reports (26), our results show that LCMV does not initially disseminate well in mouse strains lacking normal splenic architecture due to genetic mutation, such as RAG Ϫ/Ϫ , SCID, TNFR1 Ϫ/Ϫ , or LT␣ Ϫ/Ϫ (Table II and Fig. 4A). In addition to defective splenic architecture, these strains exhibit severely re-duced or absent proportions of MZMs, MMs, and MAdCAM-1 ϩ sinus-lining cells (27)(28)(29). Thus, the perturbed structure of the spleen, or possibly the absence of particular viral targets, must not allow early replication and spread to occur. In contrast, the mice depleted of MZ constituents by clodronate liposome treatment maintain an otherwise normal splenic architecture. Under these conditions, early viral replication with dissemination to other areas has been reported to occur (30,31), and is also apparent in the experiments presented here (Fig. 4, A and B). Taken together, these observations indicate that although LCMV exploits the normal architecture to infect the first cells it encounters in the spleen, the host requires the presence of these cells to confine replication and prevent dissemination. In this respect, the MZ is an indispensable region for capture of LCMV.
By extending the analysis to type 1 IFN production as well as viral dissemination, our studies reveal an ability of the host to contain LCMV through the reconstitution of particular MZ subsets before restoration of a type 1 IFN response. Therefore, two important splenic functions are identified and separated. Interestingly, clodronate liposome-sensitive MZ cells are also important for controlling MCMV replication (18) (Fig. 2C). In contrast to LCMV, however, this can be demonstrated despite normal type 1 IFN responses (Fig. 2, C and D). Thus, cellular contribution to the containment of replication can be observed in two different viral systems independently of the presence or absence of the type 1 IFN response.
The results suggest that conclusions from earlier reports regarding the role of the MZ in promoting defense against infection may need to be revisited. LCMV infection of either clodronate liposome-treated mice or mice lacking proper splenic architecture results in reduced CD8 T cell activity (26, 30 -33), and this has been attributed to high antigenic load (in the case of clodronate liposome-treated mice) or suboptimal priming of CTLs (in the case of mice lacking proper architecture). As the type 1 IFNs are well documented, however, to be involved in several aspects of CD8 T cell priming, proliferation, and survival (4, 5, 34, 35), the absence of the cytokine response in clodronate liposome-treated mice or . Mice with aberrant splenic architecture control LCMV replication whereas clodronate liposome-treated mice exhibit viral dissemination, which is controlled with the reappearance of CD11c ϩ cells. (A) C57BL/6 mice (f) or (as indicated, Ⅺ) SCID mice, TNF-R1 Ϫ/Ϫ mice, or mice treated with liposomes at day Ϫ1 and then infected at day 0 (d0) or at day 14 (d14) were infected with LCMV and harvested at day 2 following infection. Plaque assays were performed on serum or spleen homogenate as described in Materials and Methods. B, C57BL/6 mice were uninjected (f) or i.v. injected with clodronate liposomes (Ⅺ) at day Ϫ1 then uninfected (UN) or infected i.p. with LCMV at days 0, 7, or 14 following liposome administration. All mice were harvested at day 2 following infection. LCMV viral titers were determined on serum and spleen homogenates by plaque assay. C, Immunohistochemical analysis from the control or liposome-injected mice analyzed in B. Immunohistochemistry was performed as in Fig. 3B. Results are presented as means Ϯ SEM of two to five mice per group and representative of at least two independent experiments. , Below limit of detection. mice lacking splenic organization may also contribute to reduced CD8 T cell activity. Type 1 IFNs have been documented to promote the phenotypic and functional activation of DCs (2, 36 -39), and recent reports have shown that LCMV-specific CD8 T cell priming is dependent on the presence of MZ cells (40,41). Thus, in clodronate liposome-treated mice or mice lacking proper architecture, MZ cells (DCs, in particular) may not become activated to induce CD8 T cell priming due to the absence of type 1 IFN.
In summary, the findings presented here demonstrate that splenic architecture, particularly of the MZ, is important for inducing the production of type 1 IFN as well as controlling viral replication during infections with particular viruses. The results significantly advance our understanding of the role of the spleen in defense against blood-borne pathogens and in the activation of endogenous innate and adaptive immune components. Given that some individuals are missing the organ in its entirety and there are defects in splenic composition under a variety of genetic conditions, including those observed in SCID patients, the observations have major implications concerning the different capacities and characteristics of endogenous immune responses. They suggest that people with defects in splenic composition may be less able to produce type 1 IFN against certain viruses, but also offer the possibility of exogenous administration of the cytokines to help control these pathogens. Taken together, this information advances the understanding of the many factors that can determine resistance or susceptibility of individuals to particular infectious organisms.

Disclosures
The authors have no financial conflict of interest.