Alterations in Marginal Zone Macrophages and Marginal Zone B Cells in Old Mice Free

More than 12.9% of the U.S. population is over the age of 65 (U.S. Census Bureau, 2009 [http://quickfacts.census.gov/qfd/states/00000.html]) and life expectancy continues to make rapid gains as a result of advancements in modern healthcare. However, older people suffer greater risks of long-term complications and are still more susceptible to diseases than young and middle-aged individuals because of the effects of a weakened immune system. One of the leading causes of death in persons aged 65 years and older is invasive pneumococcal disease (1, 2). Vaccines for the prevention of pneumococcal disease show a decline in immune protection in the elderly when compared with the young (3, 4). This finding is possibly caused by the reduced ability of the aged immune system to deal with bacteria load and mount an immune response to the T-independent (TI) Ag components that compose the vaccine. Therefore, understanding how the immune system changes with age is critical for implementing better therapies and vaccines for the prevention of age related disease.

Blood-borne Ags are detected by in vivo imaging to initially enter the spleen in relation to other immune organs (5). Blood-borne bacteria, viruses, parasites, and other Ags enter a compartmentalized area of the spleen, the marginal zone (MZ), where they are sequestered by specialized MZ macrophages (MZMs) and MZ B cells (6, 7). MZMs are highly phagocytic cells and are responsible for rapid clearance of blood-borne TI Ags and debris (8–10). MZ B cells are also well known for their ability to respond to TI Ags by rapidly generating an Ab response (11–13). Notably, the Ag components of vaccines important for the prevention of bacterial pneumonia are TI (14). Thus, examining the MZ compartment, especially MZMs and MZ B cells, in older patients may provide an explanation for increased susceptibility and decreased efficacy of vaccines for bacterial diseases.

MZMs are capable of binding TI Ags through specific cell surface receptors. Two important MZM cell surface receptors are: macrophage receptor with collagenous structure (MARCO, a scavenger receptor) and specific intracellular adhesion molecule-grabbing nonintegrin receptor 1 (SIGN-R1) (7). MARCO binds to Staphylococcus aureus (15, 16) and Escherichia coli (16); SIGN-R1 (a homolog of human DC-SIGN) (17) binds the capsular polysaccharide of Streptococcus pneumoniae and also to the polysaccharide dextran (9, 18–20). Once MZMs bind Ag, they introduce it to closely associated MZ B cells (21). MZMs and MZ B cells have a direct intercellular interaction via MARCO expressed on MZMs with an undetermined ligand on MZ B cells (22–24). MZMs and MZ B cells are positioned around the outer border of the MZ sinus facing the red pulp of the spleen (7). Positioned around the inner border of the MZ sinus are the metallophilic macrophages (MMMs), which face the white pulp of the spleen (7). The anatomic microarchitecture of the MZ can be seen as critical for the proper binding and clearance of blood-borne pathogens.

Previous studies of the lymphoid architecture of old mice focused on disruption of the white pulp areas, T cell-rich periarteriolar lymphoid sheath, and B cell follicles, when compared with young mice (25–29). Because little is known about changes in the structure and cellular components of the MZ with age, the purpose of this study was to compare this important immune compartment in old and young mice. We determined that the splenic tissue of most old mice have an overall alteration in the microarchitecture of the MZ, including decreased frequencies of MZMs and MZ B cells and disruptions in positioning of the mucosal addressin cell adhesion molecule 1 (MAdCAM-1)⁺ sinus lining cells and MMMs. The reduction of MZMs was particularly striking, and in individual aged animals the reduced abundance of MZMs in old mice correlated with decreased frequency of MZ B cells. The consequence of altered MZMs in old mice was reflected in reduced binding of blood-borne dextran particles by the MZMs. However, the phagocytic function of individual MZMs was found to be unaffected by age, suggesting that a decrease in the MZM compartment rather than cell intrinsic defects may be a key factor in the decreased immune response to TI Ags reported with age.

Female BALB/c mice were purchased from Harlan Laboratories (Indianapolis, IN) or Charles River (Wilmington, MA) through a contract with the National Institute of Aging at 17–18 mo of age and from Harlan Laboratories at 6–8 wk of age. Upon receipt, the animals were housed at the animal research facility (Loyola University Medical Center, Maywood, IL) under specific pathogen-free conditions until being sacrificed. Old animals with any visual abnormalities such as enlarged spleens and presence of tumors were not used in these studies.

For H&E staining, spleens were fixed in 10% buffered formalin and processed using an automatic tissue processor, dehydrated through graded alcohols, cleared in xylene, and infiltrated with paraffin. Paraffin sections were sectioned on a microtome (4 μm) onto Superfrost Plus microscope slides (Fisher Scientific, Pittsburg, PA). Sections were then treated with xylene to deparaffinize the tissue and then rehydrated in successive decreasing grades of alcohol. The splenic sections were stained with H&E. Tissues were viewed and photographed with a Leitz Diaplan microscope (W. Nuhsbaum, McHenry, IL). For frozen sections, one third of each spleen was flash-frozen in Tissue-Tek O.C.T. freezing medium (Sakura, Torrance, CA) and stored at −80°C. Frozen spleens were cryosectioned (6–8 μm) onto Superfrost Plus microscope slides. Splenic sections were fixed in acetone for 10 min prior to immunocytochemistry or immunofluorescence staining protocols.

H&E-stained slides were blind labeled and scored independently by two investigators. In consultation with a pathologist, criteria were established to evaluate the morphology of the MZ. Four well-oriented white pulp areas were chosen for each animal. First, the interface between the white pulp and MZ was evaluated subjectively for the percent of distortion around each white pulp area and given a score ranging from 0 to 4 (Fig. 1A). A score of 0 was ≥75% distortion; 1 corresponded to ≥50% and ≤75%; 2 corresponded to ≥12.5% and ≤50%; 3 corresponded to ≤12.5%; and 4 corresponded to no evidence of interface distortion. Second, each white pulp area was evaluated for the extent of interface distortion by determining the percent of radius involvement, as defined by the amount of distortion that protruded toward the center of the white pulp. As before, for each white pulp area, a score ranging from 0 to 4 was given. A score of 0 corresponded to ≤20% distortion; 1 corresponded to ≤15%; 2 corresponded to ≤10%; 3 corresponded to ≤5%; and 4 corresponded to no distortion protruding inwards. The summation of the interface distortion and percent of radius involvement per white pulp area was determined and given a total score ranging from 0 to 8. The final distortion was characterized as severe (0–2), moderate (3 or 4), mild (5 or 6), or minimal (7 or 8). Finally, the average score of all four white pulp areas per animal was determined and reported for each mouse.

Frozen sections were fixed and then incubated with 5% normal mouse serum to block nonspecific protein interactions (The Jackson Laboratory, Bar Harbor, ME). The sections were then incubated for 1 h with an unconjugated rat anti-mouse IgG2a anti–MAdCAM-1 mAbs (MECA-367; BD Pharmingen, San Diego, CA) followed by a 30-min incubation with alkaline phosphatase conjugated mouse anti-rat IgG2a (2A 8F4; Southern Biotech, Birmingham, AL) to detect the MZ sinus lining cells. Positive cells were revealed using the Alkaline Phosphatase Substrate Kit III (Vector Laboratories, Burlingame, CA). For double immunocytochemistry to detect MZMs, sections were then blocked with 5% normal mouse serum for 20 min and subsequently incubated for 30 min with either unconjugated rat anti-mouse IgG1 anti-MARCO mAbs (ED31; BMA Biomedicals, Augst, Switzerland) or unconjugated Armenian hamster anti-mouse SIGN-R1 (ebio22D1; eBioscience, San Diego, CA) followed by a 20-min incubation with HRP-conjugated mouse anti-rat IgG1 (G1 1C5; Southern Biotech) or rabbit anti-Armenian hamster IgG HRP (Abcam, Cambridge, MA), respectively. Positive cells were revealed using the DAB Substrate Kit for Peroxidase (Vector Laboratories). All Ab incubations were at room temperature. Slides were counterstained with Nuclear Fast Red (Vector Laboratories) for 60 s, mounted with VectaMount (Vector Laboratories), and viewed and photographed with a Leitz Diaplan microscope.

MARCO was revealed by immunocytochemical staining of spleen frozen sections. Slides were blinded, and four well-oriented white pulp areas were chosen per animal and scored independently by two investigators. Each area was given a score of 1, 2, or 3 for the abundance of MARCO⁺ MZM and the percent of the continuity of MZM encircling each white pulp area. A score of 3 equated to a high abundance of MZMs present that tightly encircled the MZ sinus. A score of 2 equated to a moderate abundance of MARCO⁺ MZM that partially encircled the MZ sinus. A score of 1 equated to low or virtual absence of MARCO⁺ MZM, that when present, were found in patches along the MZ sinus. An average score of the four white pulp areas graded was determined and reported for each individual animal.

Cryosections of spleen were fixed and incubated with Super Block (ScyTek Laboratories, Cache, UT) for 5 min according to manufacturer’s protocol. Sections were incubated at room temperature with primary Abs for 1 h followed by respective secondary Abs for ∼30–45 min. To identify MZ sinus lining cells, sections were stained with unconjugated rat anti-mouse MAdCAM-1 or unlabeled rat IgG2a followed by goat anti-rat Alexa Fluor 633 (Invitrogen, Carlsbad, CA). To reveal MMM, sections were incubated with FITC conjugated rat anti-mouse CD169 (MOMA-1; AbD Serotec, Oxford, U.K.) or with rat IgG2a FITC. To reveal MZMs, sections were stained with unconjugated Armenian hamster anti-mouse SIGN-R1 or with Armenian hamster IgG, followed by goat anti-hamster IgG biotin (The Jackson Laboratory) and Streptavidin-Dylight 549 (The Jackson Laboratory). Sections were blocked with 5% normal Armenian hamster serum for 20 min and stained for colocalization of MARCO using unconjugated rat anti-mouse MARCO or with isotype control rat IgG1, followed by donkey anti-rat IgG Dylight 488 (The Jackson Laboratory). To reveal MZ B cells, mice were injected i.v. in the tail vein with 1 μg PE-conjugated rat anti-mouse CD21/35 (8D9; eBioscience) in 200 μl sterile PBS, according to Cinamon et al. (30). The mice were sacrificed after 5 min, and one-third spleen was placed in Tissue-Tek O.C.T. freezing medium and cryosectioned as described above. Sections were costained with rat anti-mouse MAdCAM-1 biotin (MECA-89; BD Pharmingen) followed by Streptavidin Alexa Fluor 488. Slides were mounted with fluorescent mounting medium (Dako, Carpinteria, CA) and viewed using a Zeiss LSM 510 confocal microscope (Carl Zeiss, Jena, Germany).

MZ sinus lining cells were revealed by immunofluorescence staining of frozen spleen sections from the middle third of the spleen. In young animals, the MAdCAM-1 stain forms a thin line outlining the outer white pulp. To compare staining in spleens from aged mice, a tiled image (original magnification ×25) of a complete cross section of spleen (6–8 μm) from each young or aged animal was compiled using a Zeiss LSM 510 confocal microscope (Carl Zeiss). Using LSM image browser software and the polyline drawing and measurement tools, lines were drawn along the MAdCAM-1⁺ cells surrounding each white pulp area, and measurements of the total length of MAdCAM-1⁺ cells enveloping all white pulp areas in a single section were recorded. Disrupted lengths were then discriminated and the length of the affected area was measured by identifying areas of dispersed, noncompact MAdCAM-1⁺ staining. Relative proportions of the disrupted areas versus the total length of the lines covered by MAdCAM-1⁺ cells were calculated.

Using a protocol previously described by Ato et al. (31), 200 μl dextran-FITC (500,000 m.w. Invitrogen) was injected i.v. in the tail vein of mice (100 μg/ml) in sterile PBS. After 45 min, the mice were sacrificed and spleens were harvested. One third of each spleen was placed in Tissue-Tek O.C.T. freezing medium for immunofluorescence analysis, and the remaining 2/3 was used for flow cytometric analysis (described below). For colocalization of MARCO with dextran-FITC positive cells, cryosections were incubated with unconjugated rat anti-mouse MARCO followed by donkey anti-rat IgG Dylight 488 as described above. Slides were mounted with fluorescent mounting medium and viewed using a Zeiss LSM 510 confocal microscope.

Young and old mice were injected with dextran-FITC as described above. Splenocytes were harvested and 4 × 10⁶ cells in 1 ml/well were allowed to adhere onto four-well chamber slides (Nunc Labtek) for 1 h in 37°C incubator. Nonadherent cells were removed and pHrodo S. aureus BioParticles (pHrodo; Molecular Probes Invitrogen) were added at 100 μg/ml in HBSS (Life Technologies Invitrogen) containing 20 mM HEPES (Life Technologies Invitrogen) for 1 h in an incubator at 37°C. Cells were then viewed using a Leica TCS SP5 two-photon confocal microscope (Leica Microsystems) equipped with a heated chamber at 37°C. For each animal, dextran⁺ pHrodo⁺ MZMs were imaged live (original magnification ×5000). Three criteria were needed to qualify as a MZM: first, cells were dextran⁺; second, cells had macrophage morphology; and third, cells were pHrodo⁺. Images were made at approximately the center of the cell. Phagocytosis of pHrodo by MZMs was semiquantitated using ImageJ version 1.43u software. For each animal, individual dextran⁺ pHrodo⁺ MZMs were evaluated for the mean fluorescence intensity (MFI) of pHrodo or dextran. The relative size (area in arbitrary units) of each MZM analyzed was also determined using ImageJ software. The number of pHrodo⁺ phagocytic compartments per representative image for individual MZM was also counted. Phagocytic compartments were also counted on Z-stack images of selected cells to verify that counts using the representative images approximated the average number of compartments per image for an entire cell. For each young mouse, 7–10 MZM images were analyzed, and 4–5 MZM images were analyzed for each old mouse. Fewer MZMs were detectable in young mice.

Splenocytes were dispersed in approximately half to two thirds of each spleen by gently pressing the tissue between two frosted glass slides. The dispersed cells were collected in RPMI 1640 with 10% FCS, and small clumps were allowed to settle in a conical tube and discarded. Erythrocytes were lysed with 0.8% ammonium chloride solution in PBS and incubated on ice for 4 min. For four-color analysis of MZ B cells, 1 × 10⁶ cells per sample were incubated with rat anti-mouse CD23-PE-CY7 (B3B4; eBioscience), rat anti-mouse B220-FITC (RA3-6B2; Southern Biotech), rat anti-mouse CD21/35-PE, and rat anti-mouse IgM-PE-Cy5.5 (II/41; eBioscience) for 20 min on ice, per a protocol modified from Won et al. (32). Cells were then washed, fixed in 1% paraformaldehyde, and analyzed using FACS Canto II with Tree Star FlowJo software version 7.5.5 (Becton Dickinson, Mountain View, CA). MZ B cells were identified as CD23⁻ B220⁺ CD21/35^high IgM^high.

For MZM characterization, spleen cells were treated with Collagenase D (Roche Diagnostics, Indianapolis, IN) for 1.5 h at 37°C, and single-cell suspensions of splenocytes were recovered after gentle filtration of the minced tissue through a fine nylon mesh. The erythrocytes were lysed, and the cells were washed and resuspended in RPMI 1640. For two-color analysis to identify MZMs, 1 × 10⁶ cells per sample were pretreated on ice with rat anti-mouse CD16/32 FcR-block (FCR4G8; AbD Serotec) and subsequently incubated with either allophycocyanin-Cy7 labeled rat anti-mouse CD11b (M1/70; eBioscience) or biotinylated CD11b (M1/70; eBioscience) and Armenian hamster anti-mouse SIGN-R1-Alexa Fluor 647 (22D1; eBioscience) for 20 min on ice. Streptavidin-PerCP (16904; BD Pharmingen) was used as a second-step reagent.

Statistics were calculated using GraphPad Prism software (version 5.01). Mann–Whitney U test, Unpaired t test, and Pearson correlation were applied where appropriate.

A substantial literature attests to the high degree of morphologic organization of the MZ region in young animals (7, 33, 34). In consultation with a pathologist, two morphologic criteria were selected to discriminate the degree of morphologic disruption in the MZ areas of aged spleens. Sections of H&E-stained spleens from young and old mice were scored independently by two investigators who were blinded to the identity of the specimens (Fig. 1A). The criteria used were 1) the interface distortion, defined as the boundary between the white pulp area and the inner edge of the MZ, and 2) the percent radius involvement, defined as the percentage of interface distortion protruding toward the center of the white pulp. Each criterion was given a score ranging from 0 to 4 (severe to minimal). Four MZ areas were scored for each spleen. A sum of scores for both criteria was made for each MZ area, and this sum was averaged to yield a value for each individual spleen (Fig. 1B). The MZs from young mice (2–6 mo; Fig. 1A, top panels) uniformly encircled the white pulp areas. However, the MZs of old mice (18–23 mo; Fig. 1A, bottom panels) showed a great deal of variability, and many had a high degree of distortion around the white pulp compared with spleens from young mice. In many old mice, the splenic MZ areas were diffuse and some were difficult to discern. Fig. 1B shows that, on average, the morphology of splenic MZs from old mice could be histologically discriminated from those of young mice. Of note, the splenic MZs from some old mice retained a relatively young-like morphology, in which the MZ areas appear to have intact organization by the criteria assessed.

Because the MZ is a compartment composed of different macrophage types and B cells that are highly organized for the function of detecting and clearing Ags from the blood, our goal was to determine which cells, if any, or cellular positioning within the MZ changed in old age. Staining was included for MAdCAM-1⁺ MZ sinus lining cells to provide a landmark, because both MZMs and MMMs relate in position to the MZ sinus. Similar to numerous reports of spleens from young mice, MAdCAM-1⁺ stain (purple or red) revealed continuous, thin rings of MZ sinus lining cells surrounding white pulp areas (Fig. 2A, 2B, top panels). In contrast, the MAdCAM-1⁺ cells exhibited a less even, more diffused distribution, and they often appeared as multiple layers of cells oriented toward the white pulp in spleens from old mice (Fig. 2A, 2B, bottom panels). In spleen sections from old mice, ∼90% of the linear portion of the MAdCAM-1⁺ stain appeared visually altered compared with MAdCAM-1⁺ stain in spleens from young mice (Fig. 2A, far right panel; n = 5 for both age groups, p = 0.0001). In spleens from young mice, MMMs (green) were closely aligned with the MAdCAM-1⁺ MZ sinus lining cells and formed a continuous line of cells on the follicular side of the MZ sinus (Fig. 2B, middle top panel), consistent with published reports (7). Again, in the spleens from old mice contrasted with young, the MMMs were aligned along the MZ sinus but were highly diffused into the white pulp-follicle area (Fig. 2B, middle bottom panel). The diffused distributions of MAdCAM-1⁺ MZ sinus lining cells and MMMs were consistently observed in old mice.

Another pronounced difference between old and young spleens was in the distribution of the MZMs, marked by expression of either MARCO or SIGN-R1. MZMs in young spleens formed continuous rings on the outer edge of the MZ sinus (Fig. 2C, left and middle upper panels), as reported by others (7). However, in spleens from many old mice, MZMs appeared to be greatly reduced and aligned in discontinuous patches along the MZ sinus (Fig. 2C, left and middle lower panels). In summary, we observed several changes in the positioning of cells specific to the MZ region, including the MAdCAM-1⁺ sinus lining cells, MMMs, and MZMs. We next assessed the degree of alteration in the MZM population within individual old animals, because of the critical importance of MZM in clearance of blood-borne bacteria (7, 20) and because more variability was observed between aged individuals in disruption of MZMs than was observed in MAdCAM-1⁺ cells and MMMs.

Sections of young and old spleens were stained using Ab to MARCO (brown) to mark MZMs and Ab to MAdCAM-1 (blue) to mark the MZ sinus lining cells. Fig. 3A (top panel) shows the typical, consistent pattern observed in young spleens: the MZM population was present in a continuous line of cells adjacent to the sinus lining cells. However, the distribution of MZMs in the spleens of old mice was highly variable, ranging from a young-like array of MZMs to a negligible stain for MZMs. Many spleens from old mice had notably decreased staining for MARCO (Fig. 3A, middle panel; representative of intermediate reduction). In these spleens, the expression of MARCO was localized in patches rather than in a continuous line of cells around the MZ sinus. A semiquantitative scoring was devised to compare the MZM distribution in young versus old spleens. For each spleen, four well-oriented white pulp areas were graded blind for the relative abundance of MARCO⁺ MZMs present and the continuity of MARCO⁺ MZM positioning around the MZ sinus. Scores ranging from 1 to 3 were given for each criterion graded (worst to best; Fig. 3A shows representative tissue sections for each score). An average of scores is shown for individual animals (Fig. 3B). Expression of MARCO⁺ MZMs showed a statistically significant decrease in old tissue compared with young tissue (n = 13 young, n = 15 old, p = 0.0012). In some animals, serial sections were stained for MARCO and SIGN-R1 by immunocytochemistry to confirm that reduction of MARCO expression reflected a decreased presence of MZMs. Although less intense than for MARCO, SIGN-R1 staining reflected a similar pattern of reduced abundance and patchy stain (data not shown; example of immunofluorescent costain is shown in Fig. 2C).

To further explore the interpretation that MZMs were reduced in old mice, MZMs were quantified using flow cytometric analysis. Many reports agree that MZMs express MARCO and SIGN-R1 (9, 15, 18). Although anti-MARCO Ab worked well in immunocytochemistry, anti-SIGN-R1 staining was easier to distinguish in flow cytometry. MZMs comprise a minor subset of dispersed spleen cells; therefore, another marker was used to aid in gating the population. Published reports indicate that MZMs lack expression of the typical macrophage marker F4/80 (9) and express negligible CD11b (35). Using flow cytometric analysis, we found that the SIGN-R1⁺ population was consistently identified within a population of large cells that expressed very low levels of CD11b (Fig. 4A). The frequency of SIGN-R1⁺ MZMs of total splenocytes was then determined for individual young versus old animals. Fig. 4B shows that there was a significant decrease in the average frequency of MZMs in old mice when compared with young mice (n = 14 young mice, n = 15 old mice, and p = 0.0163). Although 60% of the old mice had MZM frequency less than that of all the young mice, we observed again that the frequency of MZMs in some old mice overlapped with the MZM frequency in young animals. In a smaller, independent study to ascertain absolute cell number, we found that young mice (n = 8) had 0.81 ± 0.10 × 10⁶ MZMs per spleen versus 0.29 ± 0.01 × 10⁶ MZMs per spleen in old mice (n = 4).

Interestingly, there was a significant increase in SIGN-R1 expression on MZMs from old mice (p = 0.0088; Fig. 4C). However, this increase in SIGN-R1 may be attributed to an increase in MZM size with age, because MZMs from old mice were found to have higher forward scatter values when compared with young MZMs (p = 0.0006; Fig. 4D). The results indicated a decline in abundance of and area covered by MZMs in the MZs of aged splenic tissue.

Because an important function of MZMs is the clearance of blood-borne Ags, we asked whether reduced binding would be a consequence of the changes in MZMs with age. Based on previous reports that dextran particles bind to SIGN-R1 and are rapidly cleared from the blood by MZMs (9, 18), experiments were performed using FITC-labeled dextran to mimic the capture of blood-borne Ag in vivo (9, 31). Dextran-FITC was administered i.v. into young and old mice. After 45 min, the spleens were frozen in OCT for visualization of dextran binding to cells in tissue sections. Binding of dextran-FITC to MZMs in spleens from both young and old mice was confirmed by costaining with anti-MARCO Ab (Fig. 5A). Spleen sections were stained with anti-MAdCAM-1 to distinguish the MZ sinus lining cells as a landmark. Confocal microscopy revealed a clear ring of bound dextran-FITC adjacent to the MAdCAM-1⁺ cells in the MZs of young spleens (Fig. 5B, upper panels). In contrast, deposition of dextran-FITC was less and in a patchy array in the MZ areas in old spleens, similar to tissue staining patterns for MARCO (Figs. 2C, 3A, 5B, lower panels).

Clearance of pathogens requires phagocytosis and binding to MZMs. MZMs are known for their highly phagocytic properties (10), and a decrease in functional MZMs could amplify reduced clearance caused by demised cell numbers in the aged spleen. Therefore, we next compared the phagocytic function of MZMs isolated from young and old mice. MZMs comprise 1–2% of total splenocytes, and isolating this relatively small population of splenocytes is difficult. We took advantage of the rapid in vivo binding of FITC-labeled dextran to SIGN-R1 on MZMs to visualize these cells after harvest and culture. Adherent splenocytes obtained from dextran-FITC injected mice were incubated with pHrodo to measure phagocytic ability of MZMs. S. aureus has been shown to bind the MARCO receptor on MZMs (16). The pHrodos are pH sensitive and emit fluorescence only after phagocytosis in the acidic conditions of the endosomal and phagolysosome compartments. Live confocal microscopy was used to visualize pHrodos phagocytosed by MZMs (Fig. 6A). Approximately 80% of dextran⁺ cells with macrophage morphology were also pHrodo positive. The MFIs of pHrodo fluorescence were determined in images of individual dextran⁺ pHrodo⁺ MZMs from both young and old mice. Within each experiment performed, MFI values for each old mouse were compared with the average value of MZMs from young animals. MZMs from old mice showed no difference in the amount of pHrodo phagocytized when compared with MZMs from young mice (Fig. 6B; n = 10 young mice, n = 9 old mice, p = 0.2689). Similarly, no differences were found between the age groups in the number of phagocytic compartments within individual cells (Fig. 6C; n = 8 young mice, n = 7 old mice, p = 0.5989). The MFI of dextran bound by young and old MZMs revealed no significant difference between age groups (Fig. 6D; n = 9 young mice, n = 9 old mice, p = 0.4256). To rule out differences in cell size of adherent MZMs in young versus old mice, which may affect the MFI, the sizes (area) of the MZMs from young and old mice were also determined, but were not significantly different between age groups (Fig. 6E; n = 10 young mice, n = 9 old mice, p = 0.1317). Notably, after imaging numerous fields from both young and old adherent splenocytes, we found that the young mice had ∼40% more dextran⁺ pHrodo⁺ MZMs compared with cultures from old mice, consistent with the flow cytometry and immunocytochemical staining showing that the frequency of MZMs declined with age. However, the MZMs remaining in aged spleens appeared to have normal phagocytic function.

It is well established that MZMs directly interact with MZ B cells via the scavenger receptor MARCO located on the surface of MZMs with an undetermined cell surface ligand on MZ B cells (22–24). Nolte et al. (36) found that in young adult mice, loss of the B cell population corresponded with loss of MZMs. Furthermore, Frasca et al. (37) reported diminished frequency of MZ B cells in old mice. Therefore, we predicted that spleens from old mice with decreased MZMs may also show a decline of the MZ B cell population in the spleen. For each animal, two thirds of the spleen was prepared for flow cytometric analysis in which both MZM and MZ B cell frequencies were determined. The remaining one third of the spleen was frozen for immunocytochemistry staining to assess the MZM qualities described in Fig. 3. Flow cytometry gating on MZ B cells was adapted from Won et al. (32) (Fig. 7B; spleen cells were initially gated on the B220⁺, CD23⁻ population followed by gating on the CD21/35^high and IgM^high fraction). Fig. 7A shows that the analysis confirmed a significant reduction in average frequency of MZ B cells in old versus young mice (n = 15 young mice, n = 17 old mice, p = 0.0002). Although the average reduction was statistically significant, a wide variation was again apparent in individual old mice, with 65% of the old mice below that of all the young mice. Some spleens showed extremely reduced MZ B cells, and other spleens had MZ B cell frequencies equivalent to the midrange of the young group. In an independent study to determine absolute cell number per spleen, young mice (n = 8) had 1.20 ± 0.14 × 10⁶ MZ B cells per spleen versus 0.52 ± 0.12 × 10⁶ per spleen in old mice (n = 4).

Moreover, the reduction in MZ B cell frequencies correlated with reductions in MZMs assessed by the semiquantitative tissue scoring method (Fig. 7C; r = 0.6429, p = 0.0003) or by frequency determined in flow cytometric analysis (Fig. 7D; r = 0.5857, p = 0.0278). Because MZMs are known to interact with MZ B cells (22–24), an assessment of the location of MZ B cells in aged tissue was done by marking MZ B cells in vivo with PE-labeled anti-CD21/35 Abs injected i.v. for 5 min based on the report by Cinamon et al. (30). Fig. 7E shows that the position of MZ B cells in old mice was generally similar to that observed in young mice.

The MZ of the spleen is well known to be the site where blood-borne Ags first enter the spleen at the MZ sinus. In this study, we describe age-related changes in the overall architecture of the MZ compartment and specific reductions in MZMs and MZ B cells. The MZs of young adult mice have a characteristic, uniform, dense appearance around the white pulp and follicles. In contrast, the MZs of old mice have a statistically significant cellular disruption when compared with young mice. Alterations to the MZ included 1) reduction in abundance of MZMs, but no differences in the phagocytic function of individual MZM, 2) reduction in the frequency of MZ B cells that correlated with the reduction observed with MZMs, and 3) unusual staining patterns for MAdCAM-1 sinus lining cells and MMMs that may indicate positional changes.

An age-related decrease in MZMs may have consequences on the immunity in aged individuals. If the spleen architecture in humans suffers a similar decline with advanced age, one consequence may be less efficient clearing and response to bacteria, as has been reported in patients with asplenia or after splenectomy (38). For example, pneumonia caused by S. pneumoniae is one of the leading causes of mortality in advanced age (1, 2). SIGN-R1 [the murine homolog of human DC-SIGN (17)] on MZM binds the capsular polysaccharide of S. pneumoniae (19, 20). It is possible that a decrease in the SIGN-R1⁺ MZM pool may be a factor for increased susceptibility to S. pneumoniae reported in older patients. To determine whether reduced frequency of MZMs affected Ag clearance in the MZ, fluorescently labeled polysaccharide dextran, which binds SIGN-R1 (9, 18), was administered via the tail vein into young and old mice. On tissue sections, less dextran was observed in the MZs of most old mice compared with young mice (Fig. 5B). When we examined the phagocytic function of individual MZMs, we found no differences between young and old mice, implying that it is the reduction in the MZM population rather than loss of cell function with age that may contribute to the age-related risk for certain pathogenic infections.

Little is known about the homeostasis of the MZM compartment in the adult. Postnatally, maturation of the MZ compartment, including establishment of the MZM layer, is delayed compared with other lymphoid areas of the spleen (39, 40). This delay is thought to be the underlying cause for increased susceptibility of infants to pneumococcal disease (2, 41). Thus, both the very young and the old share two elements: compromised anatomy of the MZ and higher risk for bacterial and viral diseases. In advanced age, three possible mechanisms can contribute to the reductions in MZMs: decreased MZ B cells, decreased chemotactic factors important for MZM homing to the MZ, or decreased MZM turnover. Evidence in this study supports the first mechanism, because decreases in MZ B cells in old mice, when present, correlate with decreased abundance of MZM. Similar correlations have been noted in other studies demonstrating that gradual loss of B cells in young mice with a conditional knockout of CD79 led to subsequent loss of MZMs (36). In addition, a report by Kraal et al. (42) observed that recovery of MZM correlated with full restoration of the B cell population in the MZs of mice after in vivo liposome treatment that led to depletion of all splenic macrophages and most MZ B cells. It is also possible that the disruptions in the microarchitecture of the spleen in old tissue are caused by decreases in the chemotactic factors important for the proper homing of MZMs to the MZ. CCL21 is a chemokine secreted by gp38⁺stromal cells that is largely responsible for MZM homing to the MZ (31). Preliminary work from our laboratory found no differences in total splenic expression of CCL21 mRNA in old versus young mice; however, examination of CCL21 using anti-CCL21 Ab on splenic sections revealed reduced chemokine present along the outer border of the MZ in old tissue compared with young tissue (data not shown). Finally, it is possible that MZMs may turnover slowly, die, and not be replaced. Kraal et al. (42) showed that new MZMs will replenish a macrophage-depleted spleen by 30 d, implying that MZM precursors exist in young animals. Therefore, it is possible that with advanced age, there is a decrease in MZM replenishment, precursors, or both. It is also conceivable that a lifetime of recurring exposure and uptake of viruses and bacteria causes repeated production of proinflammatory cytokines and chemokines, leading to overall damage to the microenvironment and decrease in immune cells such as MZMs. For example, Khanna et al. (43) recently found that TNF-α is responsible for the decline in MZMs that they observed during response to heat-killed Listeria monocytogenes.

The decrease in MZ B cell frequency in old mice is puzzling because an age-related increase toward maintenance of existing MZ B cells has been predicted by Allman and Miller (44). Moreover, certain conditions in old mice are more aligned with MZ B cell development than with follicular B cell development, such as low levels of E2A and IL-7 (12, 44–51). A shift toward expanded MZ B cell differentiation at the expense of follicular B cell production with age is an attractive hypothesis, and increased MZ B cell numbers have been reported in aged B10.D2 mice (52). However, our findings confirm a report that that MZ B cells are decreased in most old mice (37). Furthermore, it may be possible that MZ B cells are being produced at expanded levels in the old mice; however, because of the collapse in the MZ microarchitecture, the newly formed MZ B cells might not home effectively to the aged MZ compartment. In earlier work, young MZ B cells transferred to old mice appeared compromised in homing to the spleen (27). Another consideration is suggested by recent studies showing that MZ B cells are lost during chronic inflammation (53, 54). Traggiai et al. (53) reported changes in B-lineage subpopulations that appear similar to those reported in old mice (55, 56). Chronic inflammation has been viewed by many investigators as part of immunosenescence (57, 58), raising the possibility that increased inflammation with age promotes gradual loss of MZ B cells and perhaps subsequent decline of MZMs.

The MZs in old mice also showed alterations in the MAdCAM-1⁺ MZ sinus lining cells, which are part of the reticular family of cells (40, 59). Moreover, these cells act as a polarized barrier to overt penetration of Ag from the MZ sinus into the follicles, whereas Ag readily moves from the sinus outward toward the MZM and MZ B cell layer (30, 60). In aged spleens, we found that the expression of MAdCAM-1 was diffused into the white pulp and less organized in old mice compared with young mice (MMMs were found to be similarly disrupted). Strikingly similar morphologic disruptions in the MAdCAM-1⁺ sinus lining cells and MMMs were reported by Girkontaite et al. (61) in sphingosine-1-phosphate receptor 3 (S1P₃) deficient mice, leading to the conclusion that S1P₃ is necessary for the positioning of MAdCAM-1⁺ sinus lining cells and MMMs along the continuity of the MZ sinus (61). S1P has also been implicated in the induction of proteins found in junctional complexes, and at least one receptor (S1P₁) was found on the surface of reticular stromal cells, as well as endothelial cells, in lymph nodes (62). Recently, Katakai et al. (40) provided evidence that the MAdCAM-1⁺ sinus lining cells derive from a type of reticular stromal cell that requires constant LTβR stimulation for survival and production of CXCL13. These observations, with the previous finding that MZ sinus lining cells of aged spleens had less CXCL13 (27), could indicate a problem in LTβR stimulation. It is possible that either S1P and its receptors or LT/LTβR provide a link between reduced MZM-MZ B cells and disposition of MAdCAM-1⁺ cells.

Understanding the alterations that occur with age in lymphoid tissues such as the spleen is important for developing more efficient therapies for preventing diseases, such as bacterial pneumonia, that have shown to be highly detrimental in the elderly (1, 2). The cells that contribute to the framework of the MZ of the spleen are largely responsible for initiating an effective immune response to S. pneumoniae and other TI Ags. In most old mice, MZs appeared disrupted by reduction or positional changes in the cellular components that would be expected to provide a front-line defense to pathogens. If viable markers could be found to allow predetermination of old versus young phenotypes, future studies may provide insight into vaccines that can target deficiencies in MZM and MZ B cell populations in aging individuals.

We thank Dr. Sherri Yong (Department of Pathology, Loyola University Chicago Medical Center) for assistance with tissue analysis, Patricia Simms and the Loyola FACS Core Facility for assistance with flow cytometry, Dr. James Sinacore for statistical help, Dr. Seth Robia and Daniel Blackwell for assistance with phagocytosis imaging, and Dr. Heather Minges Wols and Dr. Phong Le for discussions.

The authors have no financial conflicts of interest.