Splenectomy of rats selectively reduces lymphocyte function-associated antigen 1 and intercellular adhesion molecule 1 expression on B-cell subsets in blood and lymph nodes

Antigens circulating in the blood are retained in the spleen and presented to a huge number of B and T cells migrating through its different compartments.1,2 On removal of the spleen the incidence of severe infections increases both in children and adults,3,4 and there are changes in many immunologic parameters.5 One lasting change is a high increase in the number of B lymphocytes in the blood, an effect that is constantly observed both in humans6 and in experimental animals,7,8 and that comparably affects the 2 compartments of the blood, the peripheral and the marginal pools.9 

Because it is not known whether all or only certain blood B-cell subsets increase in number after the loss of the spleen, the numbers of naı̈ve (IgD⁺IgM⁺), memory (IgD⁻IgM^high), newly formed (NF; IgM^highCD90^high), early recirculating follicular (ERF; IgM^lowCD90^high), recirculating follicular (RF; IgM^lowCD90⁻), and marginal zone (MZ; IgM^highCD90⁻) phenotype B cells10 were determined in blood, lymph nodes, and bone marrow of control and splenectomized animals. In principle, the splenectomy-induced increase in blood B-cell numbers could be due to an altered migration pattern or to increased proliferation of B cells caused, for example, by a latent infection.

Because migration and proliferation of B cells are regulated by surface molecules, our hypothesis was that their expression might be altered by splenectomy. Therefore, the expression of molecules involved in initial adhesion of lymphocytes to the endothelium, in transmigration through the endothelium and the tissue, and in B-cell activation within the tissue was analyzed in all B-cell subsets in blood, lymph nodes, and bone marrow (α₄-integrins, CD44, L-selectin, lymphocyte function-associated antigen1 [LFA-1], intercellular adhesion molecule 1 [ICAM-1], major histocompatability complex [MHC] class II, interleukin [IL]-2 receptor-α [IL2-Rα] chain11-14). The functional consequences of possible changes in surface molecule expression were tested in vitro by adhesion assays and in vivo by following the traffic of labeled B cells from blood to thoracic duct lymph. In addition, the number of proliferating B cells was determined in the bone marrow, in the B-cell area of lymph nodes, and in the blood.

Our data show that loss of the spleen leads to a decrease in LFA-1 and ICAM-1 expression of B cells but not T cells in blood and lymph nodes. This accelerates the migration of B cells through lymph nodes and results in higher numbers of B cells in the thoracic duct lymph and, subsequently, in the blood.

Normal rats (LEW/Ztm), rats with high numbers of B cells (BH.1L/Won15), and rats lacking T cells (LEW/Ztm-rnu16) were obtained from the Central Animal Facility at the Medical School of Hannover and were barrier-maintained in a room with controlled environment (22 ± 2°C, 55% ± 5% relative humidity, 12-hour light/dark cycle).

Splenectomy was performed as described previously.3,7 The animals were kept under barrier-maintained conditions for 3 months before analysis to allow recovery from the operation and adjustment to the asplenic situation. Sham splenectomy was performed as described above with the exception that after cutting the splenic ligaments the spleen was put back into the abdominal cavity. Routine microbiologic monitoring according to Federation of European Laboratory Animal Science Associations (FELASA) recommendations17 was conducted on animals from both groups. They did not reveal any evidence of infection with common rat pathogens except for Pasteurella pneumotropica, which was found in both control and splenectomized rats. To further rule out infections due to the absence of the spleen, Northern blot analysis of α₂-macroglobulin was performed according to standard procedures using liver tissue. Total RNA was isolated by the guanidium isothiocyanate method.18 Total RNA (15 μg) was analyzed through a 1% 3-(N-morpholino)propane sulfonic acid (Sigma, Munich, Germany) agarose gel, followed by transfer to Gene Screen Plus (NEN, Boston, MA). The α₂-macroglobulin and β-actin complementary DNA (cDNA) probe was labeled with αCTP32 according to Random priming (Boehringer Mannheim, Mannheim, Germany). The hybridization procedure was performed as described before.18 The ratio of α₂-macroglobulin to β-actin expression was set to 1. Changes in the expression level of α₂-macroglobulin in splenectomized rats were calculated as fold activation compared to control animals. If inflammation would have been present due to loss of the spleen, the α₂-macroglobulin expression level would have been over a 100 times higher in splenectomized rats than in control animals.18 However, the level ranged between 0.4 and 2.53 for both groups.

Blood was collected from the abdominal aorta under ether anesthesia. Mesenteric lymph nodes and both femora were removed and cell suspensions were prepared as described.19 The volume and the total number of leukocytes (Coulter Counter, Luton, United Kingdom) were determined for the different cell suspensions. The B-cell subsets were analyzed by flow cytometry (FACScan, Becton Dickinson, Mountain View, CA) as described below.

Blood, lymph node, bone marrow, and thoracic duct lymph cell suspensions were washed and adjusted to 1 × 10⁶cells/well. The different B-cell subsets were identified using a biotinylated antibody against IgM,20 which was revealed with streptavidin-conjugated Red 670 (FL3; Gibco, Gaithersburg, MD) in combination with fluorescein isothiocyanate (FITC)–conjugated monoclonal antibody Ox7 against CD9021 (FL1) and a monoclonal antibody against IgD,20 which was revealed using a phycoerythrin (PE)–conjugated secondary antibody (FL2; Becton Dickinson). By staining first for FL2 and then for FL1 and FL3, no cross-reaction occurred between the PE-conjugated secondary antibody (antimouse) and the subsequently applied anti-CD90 antibody (both generated in the mouse). Six B-cell populations were identified: naı̈ve (IgD⁺IgM⁺), memory (IgD⁻IgM^high), NF (IgM^highCD90^high), ERF (IgM^lowCD90^high), RF (IgM^lowCD90⁻), and MZ phenotype B cells (IgM^highCD90⁻).10 The incubation period was 30 minutes at 4°C for each antibody. Using a FACScan and PC-LYSYS software, about 1 × 10⁵ viable cells were analyzed and the percentage of the respective population among B cells was determined.

Cell suspensions were treated as described above. In the following experiments 2 colors (FL1 and FL3) were used to define either B-cell subpopulations (NF, ERF, RF, and MZ phenotype) or T-cell subsets (CD4, CD8), and the third color was used to determine the surface molecule expression. In brief, cells were incubated with antibodies against ICAM-1 (1A29), LFA-1 (WT.1), CD44 (Ox49), α₄-integrin (HP2/1), L-selectin (HRL-3 and Ox85), IL-2 receptor (Ox39), and MHC class II (Ox6).22 The antibodies were revealed with a PE-conjugated antimouse Ig secondary antibody (FL2). Because one of the anti–L-selectin antibodies (HRL-3) is produced in hamster it was revealed using a PE-conjugated antihamster antibody (Caltag, Burlingame, CA). Next, the cells were incubated with an FITC-conjugated monoclonal antibody against CD90 (Ox7; FL1) and a biotinylated antibody against IgM (FL3) to identify the B-cell subsets, or the cells were incubated with an FITC-conjugated monoclonal antibody against the T-cell receptor (R73; FL1) and a biotinylated antibody against CD4 (FL3) to identify the T-cell subsets. Finally, the biotinylated antibody was revealed with streptavidin-Red 670 (FL3). All incubations were performed for 30 minutes at 4°C. An irrelevant mouse antibody was used as a control. To compare the surface molecule expression, blood, lymph node, and bone marrow cells of control and splenectomized animals were always analyzed on the same day, and the different suspensions were treated identically with respect to the applied suspension medium, erythrocyte lysis, and setting of FACS parameters. Using a FACScan and the PC-LYSYS software the surface molecule expression was analyzed by gating on the respective population. Forward and side scatter analysis revealed that the size of the different B-cell subsets did not change after splenectomy. This allowed direct comparison of the fluorescence intensities of the various surface molecules between the subsets of both groups.

In the in vitro adhesion assay the binding of blood B cells of control and splenectomized rats to high endothelial venules (HEVs) of various tissues was compared to that of a standard population. The assay was performed as described previously.23,24 Blood lymphocytes were obtained from normal and splenectomized animals. Standard lymphocytes from mesenteric lymph nodes of normal rats were labeled with FITC. Lymph node cryostat sections were overlaid with 100 μL of a suspension containing 1.5 × 10⁶ lymphocytes (mixture of blood lymphocytes and standard cells). Lymph node sections from control animals were overlaid with blood cells from control rats and lymph node sections from splenectomized animals were overlaid with blood cells from splenectomized rats. The slides were agitated during incubation. For immunohistochemical staining the sections were fixed in glutaraldehyde 0.2% for 10 minutes, 1% paraformaldehyde or 30 minutes, and acetone for 2 minutes.

Blood B cells were identified by alkaline phosphatase labeling of surface antigens, and the FITC-labeled standard cells were revealed by an anti-FITC antibody and peroxidase staining described above. The incubation with antibodies was performed at room temperature in a humid chamber, each step lasting 30 minutes. The first incubation step was performed using a monoclonal mouse antibody for B lymphocytes (His14).25 For the second incubation step a cocktail was used, which consisted of goat antimouse IgG (Dianova, Hamburg, Germany) and rabbit anti-FITC. Finally a cocktail of alkaline phosphatase–antialkaline phosphatase (APAAP) mouse monoclonal and swine antirabbit was added to the sections. The peroxidase reaction was developed first (standard lymphocytes) and then the B cells were revealed using Fast Blue BB salt.23,24 The sections were counterstained with hemalum, washed in triethanolamine-buffered saline (TBS) and mounted in mounting medium (Dako, Hamburg, Germany). The ratio of blood B cells and standard cells bound to the high endothelial venules was determined.

The thoracic duct was cannulated and lymph was collected as described.22,26 After the operation the rats were allowed to adjust to their surroundings for 24 hours. Thoracic duct lymphocytes were labeled with 10 μg/mL FITC as described previously.27 Then, 50 × 10⁶ cells were either taken into culture to generate the ICAM-1^high–expressing B cells (for 24 hours at 37°C, 5% CO₂, in RPMI 1640 medium containing 10% fetal calf serum [FCS], 1 mmol l-glutamine, 50 IU/mL penicillin, 50 μg/mL streptomycin), or taken directly as ICAM-1^low–expressing B cells. Before injection, surface molecule expression on B cells was determined as described above. Then FITC-labeled B cells (either ICAM-1^high or ICAM-1^low) were injected, and the lymph was collected in portions every 6 hours for 144 hours. The mean transit time from blood to lymph, that is, the time it takes for 50% of the injected cells to be recovered in the thoracic duct lymph, was determined and compared between both populations. The recovery was always greater than 90% of the injected cells within 144 hours of collection. There was no difference in the recovery of B cells expressing high or low levels of ICAM-1. This suggests that differences in the mean transit time are due to the different level of ICAM-1 expression rather than to the different times the 2 populations were cultured in vitro.

Control and splenectomized animals were given 5 mg/100 g body weight 5-bromo-2-deoxyuridine (BrdU; Sigma) intravenously 12 weeks after surgery to determine the number of proliferating B cells immunohistologically.28 One hour and 24 hours after BrdU injection the animals were anesthetized and exsanguinated via the abdominal aorta. Blood, both femora, and the mesenteric lymph nodes were removed for further processing. The femora were flushed with medium and cytospots were prepared for blood and bone marrow as described.28 Both on cytospots and on cryostat sections 2 antigens were simultaneously revealed to identify proliferating B cells as described.29 B cells were identified using the appropriate antibodies (His14,25 IgM, and IgD20) and visualized in blue.30 After denaturation the BrdU⁺ cells were identified and visualized in red as described.29 

The statistical package SPSS for Windows 10.0 (SPSS, Chicago, IL) was used to calculate the means and SDs and to determine significant differences (P < .05 in the Mann-WhitneyU test or Wilcoxon matched-pairs signed-rank test).

Three months after splenectomy the number of B cells in the blood was significantly increased, whereas the number of T cells, natural killer (NK) cells, monocytes, and neutrophils was unchanged (Table1). This increase in blood B-cell numbers was found both among naı̈ve and memory B cells (Figure1A). In addition, when the B cells were further subdivided into NF, ERF, RF, and MZ phenotypes, a significant increase was observed for all subsets (Figure 1B).

Surface molecules are involved in the regulation of migration and proliferation, 2 possible causes of the observed increase in B-cell numbers. Therefore, the expression of LFA-1, ICAM-1, L-selectin, α₄-integrins, IL-2–Rα chain, CD44, and MHC class II was determined by flow cytometry on the different B-cell subsets in the blood. Three months after splenectomy no difference in the expression of the adhesion molecules analyzed was observed between the 2 groups with the exception of LFA-1. LFA-1 expression was significantly reduced on all B-cell subsets in the blood of splenectomized animals compared with control animals (Figure 2).

Compared to blood B cells, lymph node B cells in control animals showed significantly higher expression levels of ICAM-1 (Figure3A,B) and CD44 and MHC class II (data not shown). Expression of L-selectin was reduced (data not shown). Splenectomized animals also demonstrated increased levels of ICAM-1, CD44, and MHC class II and a reduced level of L-selectin in lymph node compared to blood (data not shown). However, compared to control animals the difference in ICAM-1 expression between B cells in blood and lymph nodes was much smaller in splenectomized animals. This resulted in a significantly lower ICAM-1 expression on B cells from splenectomized than from control animals (Figure 3C). These results demonstrate that the spleen is necessary for the expression of LFA-1 and ICAM-1 on B cells in the blood and lymph nodes. In contrast, the expression of LFA-1 and ICAM-1 on CD4⁺ and CD8⁺T cells did not change after splenectomy (data not shown).

After splenectomy, nude animals (genetically lacking T cells) also developed B lymphocytosis in the blood (sham-operated, 2300 ± 354; splenectomized, 4786 ± 1031; n = 6 for each group) accompanied by a significantly decreased LFA-1 expression, as well as a significantly reduced ICAM-1 expression on lymph node B cells (data not shown). This demonstrates that also in nude animals the spleen controls the regulation of LFA-1 and ICAM-1 expression of B cells in blood and lymph nodes, suggesting that T cells are not involved in this process.

Next, it was investigated whether decreased LFA-1 expression on blood B cells would impair their entry into lymph nodes. The binding of blood B cells from control and splenectomized animals to HEVs of control and splenectomized animals, respectively, was studied in an in vitro adhesion assay. As shown in Figure4, panels A and B, blood B cells and standard cells could be clearly identified. The standard lymphocytes were derived from mesenteric lymph nodes of control rats and the binding of blood B cells from either normal or splenectomized animals was related to them. This permitted a comparison of the binding of the 2 populations. The B cells of control and splenectomized rats showed comparable binding to HEVs of different lymph nodes and Peyer patches (Figure 4C). This suggests that the entry of B cells into these organs was not impaired by the splenectomy-induced decrease of LFA-1 expression.

To test whether a decrease of ICAM-1 expression affects B-cell migration in vivo, thoracic duct B cells that differ in their ICAM-1 expression were labeled. These cells did not differ in their LFA-1 or L-selectin expression (data not shown). Either ICAM-1^high(mean fluorescence intensity 581 ± 42) or ICAM-1^low(mean fluorescence intensity 524 ± 23) B cells were injected. ICAM-1^low B cells had a significantly faster mean transit time (time point when 50% of the injected cells were recovered) from blood via lymph nodes to thoracic duct lymph than ICAM-1^high B cells (41 ± 5 hours versus 58 ± 3 hours; each n = 4). This shows that the decreased ICAM-1 expression leads to faster migration of B cells through the lymph nodes.

To analyze whether the down-regulation of LFA-1 and ICAM-1 on B cells after splenectomy influences their proliferation rate, control and splenectomized rats intravenously received BrdU, which labels proliferating cells in the S phase of the cell cycle.29One hour later, the number of BrdU⁺ B cells was determined in bone marrow and lymph nodes. Proliferating bone marrow B cells (Figure 4D) and lymph node B cells (Figure 4E) could be clearly identified. The bone marrow of control and splenectomized animals contained comparable numbers of proliferating B cells (Figure 4F), and the percentage of NF B cells within the bone marrow remained constant after splenectomy (control, 6% ± 2% versus splenectomized 7% ± 3%; each n = 4). In addition, both the number of BrdU⁺ cells in lymph nodes (Figure 4G) and the percentage of BrdU⁺ cells among the blood B cells was comparable in control and splenectomized animals (control, 1.0% ± 0.4%, n = 5; splenectomized, 1.3% ± 0.4%, n = 5) 24 hours after BrdU injection.

Together, these results indicate that changes in migration properties rather than changes in proliferation cause the B lymphocytosis in the blood after splenectomy. This conclusion is further supported by the observation that B cells from BH.1L rats, which have much higher numbers of B cells than Lewis rats, also showed a marked increase in blood B-cell number after splenectomy (Table2).

Splenectomy leads to an increase in the number of blood B cells in humans, sheep, and rats.6-8 The present study extends these observations by showing that all B-cell subsets increase in the blood after splenectomy.

In addition, our study further shows that expression of LFA-1 and ICAM-1 is down-regulated on B cells after removal of the spleen, suggesting that these molecules are involved in the generation of the B lymphocytosis. This conclusion is supported by the observation that after splenectomy T cells do not show down-regulation of LFA-1 and ICAM-1, and, as already mentioned, T lymphocytosis is not observed. Thus, the questions arise (1) how the spleen influences the expression of LFA-1 and ICAM-1 on B cells in blood and lymph nodes and (2) how the reduced LFA-1 and ICAM-1 expression is involved in the generation of the B lymphocytosis in blood.

The spleen contains 4 compartments (periarteriolar lymphatic sheath [PALS], red pulp, follicles, and MZ) which, together or individually, could be the sites involved in the regulation of LFA-1 and ICAM-1 expression on B cells. Our results show that splenectomy of nude animals has the same consequences as splenectomy of normal animals (B lymphocytosis, LFA-1 and ICAM-1 down-regulation). Because nude rats have no or only very few T cells,16 this indicates that the T cells of the splenic T-cell area (PALS) are not involved in LFA-1 and ICAM-1 regulation.

Splenic transplantation experiments point toward the MZ as the most likely site of regulation. After splenectomy, splenic tissue can be implanted into the greater omentum. Following a phase of necrosis the implanted splenic tissue regenerates.3,31 Although red pulp and follicles of such splenic transplants are well developed31 and functional,2,32 a marked B lymphocytosis is still observed in these animals,2,7suggesting that red pulp and follicles are not involved in LFA-1 and ICAM-1 regulation on B cells. In contrast, the MZ is poorly developed3,31,33 and not functional as shown by the increased susceptibility to infections.33 This suggests a role of the MZ in the maintenance of LFA-1 and ICAM-1 expression on B cells in blood and lymph nodes. Two observations further support this conclusion. First, the MZ is unique to the spleen,34 and after splenectomy no substitute is available.35 Second, the MZ is poorly developed in young children up to 2 years of age. Here, the main finding is a low expression of CD21 on MZ B cells.35 Interestingly, in children of this age a B lymphocytosis is observed that is comparable to that seen after splenectomy.36,37 

It remains to be shown why only LFA-1 and ICAM-1 are affected by splenectomy, why its expression on B cells but not on T cells is modulated, and which cell types within the MZ are responsible for the maintenance of LFA-1 and ICAM-1. In addition, it is unclear whether B cells are required to migrate through the MZ or whether MZ cells produce soluble factors that influence LFA-1 and ICAM-1 expression of B cells while they are in the blood and lymph nodes. Furthermore, it should be analyzed whether the continuous presence of the MZ is necessary or whether its presence at some time of B-cell development is sufficient.38 

At first glance the cause for lymphocytosis in the blood does not seem difficult to understand. An organ is removed, which is constantly entered by many B cells,1,2 and if it is missing B cells accumulate in the blood. However, T cells also migrate in high numbers into the spleen,2 and their numbers in the blood are not increased after splenectomy (Table 1) making the “lack of space” argument unlikely.

The comparable binding of B cells from control and splenectomized animals to the HEVs of various lymph nodes and Peyer patches suggests that the entry into these organs is not affected. This is not surprising because B cells of splenectomized animals still express LFA-1 (albeit at a reduced level) and even lymphocytes completely deficient in LFA-1 are able to enter lymph nodes and Peyer patches via HEVs in numbers ranging between 20% and 70% of control cells.39 In addition, a higher number of injected lymphocytes leaves the blood of splenectomized sheep compared to control animals.8 Together these studies indicate that the entry of B cells is not impaired. However, when lymphocytes were infused into the blood of pigs 5 days after splenectomy a slower disappearance rate was noted.40 This shows that the effects of splenectomy on lymphocyte entry change over time, which is in agreement with a recent study in rats showing that B lymphocytosis develops from 4 weeks after splenectomy onward and lasts for at least 18 months.7 

Once the B cells are within the tissue, the decreased expression of LFA-1 and ICAM-1 on B cells might weaken their interactions with other cells within the lymph nodes. In turn, this might lead to a faster migration of B cells through the tissue and might result in higher numbers of B cells in the efferent lymph, consequently causing an increased re-entry of B cells into the blood. This scenario is in accordance with the observation that labeled B cells with a reduced ICAM-1 expression migrate significantly faster from blood via the lymph nodes and efferent lymphatics back into the blood (present study), and the number of B cells in the thoracic duct is doubled after splenectomy.41 Because lymph nodes contain about 40% of all B cells and the blood about 2%,42,43 small changes in the number of B cells leaving the lymph nodes are sufficient to cause large changes in the number of B cells in the blood. It remains to be determined which cell types are involved in the interactions occurring during transmigration of the lymph nodes,44 and whether and to what extent the migration through nonlymphoid organs such as lung, liver, and gut is affected by the changes in LFA-1 and ICAM-1 expression after splenectomy.

Another explanation for the B lymphocytosis could be that splenectomy increases B-cell production in the bone marrow and at other sites, such as secondary lymphoid organs. However, the number of proliferating B cells in the bone marrow, in the B-cell area of lymph nodes, and in the blood of rats (present study) and humans45 is not affected by splenectomy, arguing against an effect of splenectomy on B-cell proliferation.

In conclusion, after splenectomy LFA-1 and ICAM-1 expression on B cells in blood and lymph nodes is reduced. This shows that the expression of adhesion molecules in one organ can depend on the presence of another distant organ and demonstrates an additional level in the regulation of surface molecule expression. It is probably the MZ that regulates the expression level of LFA-1 and ICAM-1 on B cells. The down-regulation of LFA-1 and ICAM-1 probably accelerates the migration of B cells through lymph nodes and provides an explanation for the well-known B lymphocytosis after splenectomy. Thus, the overwhelming postsplenectomy infections (reviewed in Pabst et al3) may not exclusively be due to the absence of the splenic MZ per se.35,46,47They could be the result of the accelerated migration of B cells through the lymph nodes leaving less time for the cellular interactions necessary for initiation and regulation of an effective immune response.48,49 

The authors would like to thank R. Pabst and R. Schwinzer (Hannover, Germany) for their critical comments and K. Bankes, I. Dressendörfer, S. Lopez-Kostka, and F. Weidner for the technical assistance.