Maturation of B Cell Precursors Is Impaired in Thymic-Deprived Nude and Old Mice¹

Bone marrow B cell precursors pass through a series of developmental stages during their maturation into B cells (1). During the pro-B and pre-B cell stages, the germline Ig heavy and light chain gene segments begin to rearrange, respectively. The expression of the two recombination-activating genes, RAG-1³ and RAG-2, is an absolute requirement for the expression of surface Ig (sIg) molecules and the appearance of mature B cells (2). Normal B cell development had been thought to be independent of the thymus gland, T cells, or their products because the numbers of splenic B cells in nude, adult thymectomized, or αβ-TCR-deficient mice are normal (3, 4).

We previously reported that the expression of the RAG-1 gene by bone marrow B cell precursors declines with age in mice (5). Thus, the level of RAG-1 mRNA is maximal in mice between the ages of 2 and 5 mo and thereafter diminishes. Bone marrow B cell precursors from mice >10 mo of age have levels of RAG-1 mRNA that are only 10% of their peak levels. The rate of decline in the levels of RAG-1 mRNA with age occurs later but is similar to that of thymic involution, which in mice begins after 6 wk when the total thymic cell number and migration of T cells from the thymus declines (6). By 6 mo, the total number of thymic cells has dropped to ∼30% but the number of T cells released is only 5% of peak production (6). This observation, together with the low levels of RAG-1 mRNA in nude mice and the increased levels of RAG-1 mRNA in nude mice following the injection of supernatant medium from activated T cells, suggested that the thymus gland, functional T cells, or T cell products are required for the normal expression of the RAG-1 gene and the normal development of bone marrow B cell precursors (7).

We now offer evidence that a decreased number of bone marrow pre-B cells in old and nude mice and a lower percentage of pre-B cells expressing RAG is the cellular basis for the low levels of RAG mRNA (and RAG protein) in bone marrow B cell precursors. We also show that supernatant medium from activated CD8⁺ T cells induces an increase not only in RAG mRNA but also in the number of bone marrow pre-B cells. Finally, IL-16, present in the supernatant medium from activated T cells, when injected in nanogram quantities into nude or old mice increases the levels of RAG mRNA and the numbers of pre-B cells in nude and old mice.

Female C57BL/6 and BALB/c mice between 3 and 80 wk of age were obtained from the National Institutes of Health aging colony (Charles River Laboratories, Wilmington, MA). Nude mice, 6 to 8 wk old, were purchased from The Jackson Laboratory (Bar Harbor, ME). MHC class II knockout mice, C57BL/6.FDR-[KO-A^B], obtained from Taconic Farms (Germantown, NY), are CD4⁺ T cell deficient. TAP1 deficient mice, provided by A. Berns, are CD8⁺ T cell deficient (8). Mice of both transgenic strains were 6 to 8 wk old. C57BL/6 mice were used in the initial experiments on RAG and terminal deoxynucleotidyl transferase (TdT) mRNA in old mice; BALB/c mice were used for all subsequent experiments. Mice were maintained at the Cornell University Medical College under laminar flow and killed by cervical dislocation. Routine serologic assays for viral, bacterial, and parasitic pathogens were negative. All experiments were performed in accordance with protocols approved by the Animal Care and Use Committee of Cornell University Medical College.

Single spleen cell suspensions were prepared, washed three times with HBSS, resuspended in RPMI 1640–5% FCS at a concentration of 2 × 10⁶ cells/ml, and incubated on nylon wool columns incubated at 37°C After 1 h, T cells were eluted with RPMI 1640–5% FCS. For analysis of T cell subsets, nylon wool-purified T cells were incubated with rat anti-murine CD4 or CD8 Ab (Boeringer-Mannheim, Indianapolis, IN) and separated into CD4- or CD8-enriched preparations using magnetic beads (Dynal, Lake Success, NY) coated with goat anti-rat IgG Ab (9). The T cell preparations were >90% pure and were adjusted to 10⁶ cells per ml of RPMI 1640–5% FCS and incubated for 72 h at 37°C with 10 μg/ml PHA (Sigma, St. Louis, MO) added either at the beginning or end (control T cell supernatant (CTCS)) of the culture period. The activated T cell supernatant (ATCS) was collected by centrifugation and was stable for at least 3 mo when stored frozen at −80°C.

Bone marrow cells, flushed from the marrow cavity of leg bones, were collected by centrifugation and washed three times with HBSS. Where indicated, bone marrow cells were depleted of sIg⁺ cells by panning with anti-mouse Ig (Sigma) and stained with anti-mouse CD43-FITC and anti-mouse B220-phycoerythrin for 30 min at 4°C, washed three times, and analyzed using a FACScan cytometer with CellQuest software (Becton Dickinson, Mountain View, CA). Selection for lymphoid cells was on the basis of FS vs SS. The gates were chosen to exclude cell debris and granulocytes; these gates included 75% of the total bone marrow cells. Absolute numbers of bone marrow cells were determined by direct counting of the cells flushed from the bone and percentages of cells at different stages of development obtained by flow cytometry analysis. Fluorescence was measured on a log scale.

Bone marrow cells were pelleted, and RNA was isolated using the Ultraspec RNA isolation kit as specified by the manufacturer (Life Technologies, Gaithersburg, MD). cDNAs were prepared from 5 μg of total cellular RNA by mixing with 200 ng of oligo(dT), for TdT cDNA reactions, or specific primers for the RAG mRNAs as previously described (10). The PCR reactions were conducted in a final volume of 50 μl containing 1 to 5 μl of cDNA reaction, 10 mM Tris-Cl (pH 8.3), 50 mM KCl, 2 mM MgCl₂, 20 mg/ml gelatin, 0.3 mM concentrations each deoxynucleotide triphosphate, 10 μCi of [α-³²P]dCTP, 125 ng of oligonucleotide primers, and 2 units of Taq polymerase. The primer pairs used have been published: TdT (11); RAG-1 (10); RAG-2 (10); and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (12). The PCR cycle was begun at 94°C for 2 min, followed by 25 to 30 cycles of 1 min at 62°C, 1 min at 72°C, and 45 s at 94°C. Preliminary studies established that these reaction conditions yielded PCR reactions within the linear range. Aliquots of the PCR reactions were fractionated on 2.5% agarose gels, fixed in 7% TCA, dried under vacuum, visualized by exposure to x-ray film, and quantitated by densitometry.

The method of Pannetier et al. (13) was used for quantitative PCR determinations. A competitive standard plasmid was prepared by cloning RAG-1 PCR product into pCRII (Invitrogen, San Diego, CA), linearizing the resultant plasmid with AccI, and removing terminal nucleotides by partial exonuclease digestion using T4 DNA polymerase (14). The resultant digested DNA was religated, and a clone containing a plasmid with a 18-base pair deletion was selected as the standard. For the quantitative PCR reactions, a constant amount of cDNA was added to tubes containing 0.01 to 100 pg of plasmid DNA, and standard RAG-1 PCR reactions were run to 40 cycles. Run off reactions (10 cycles) were done using a ³²P-end-labeled, nested primer, 5′-GGAAAATTTGATTTGTGGGTGTTG-3′, and the products were displayed on a DNA sequencing gel. The amount of each product was determined by densitometry, and the ratio of standard to unknown was plotted against the amount of standard. The point at which the ratio of products equaled 1 was determined, and the number of cDNA molecules calculated from the known amount of indicator plasmid.

Bone marrow cells (2 × 10⁷) were first depleted of Ig⁺ and CD43⁺ cells by panning using culture dishes coated with anti-mouse Ig (Sigma) and anti-mouse CD43 (PharMingen, San Diego, CA). B220⁺ cells in the nonadherent cell fraction were selected using magnetic beads coated with anti-murine B220 (Dynal). After magnetic separation, >95% of the cells in such preparations were sIg⁻, CD43⁻, and B220⁺ pre-B cells. These cells were fixed in 4% paraformaldehyde for 10 min at 20°C, washed in HBSS, resuspended in 70% ethanol, and stored at −20°C overnight. Before staining, the cells were washed twice in HBSS and then suspended in 0.3% Triton X-100 in HBSS. Purified pre-B cells (10⁵) were incubated for 60 min with affinity-purified rabbit Ab to mouse RAG-2, kindly provided by Dr. S. Desiderio, washed three times with HBSS, and then incubated for 30 min with peroxidase-labeled anti-rabbit Ig (Sigma). The Renaissance amplification system (DuPont, Boston, MA) was used to visualize RAG-2-expressing cells with fluorescein-conjugated tyramine. At least 200 cells from each animal were counted for each determination; in general, 2 animals were used for each treatment.

Pairwise comparisons of bone marrow pre-B cell numbers in nude or old vs young euthymic mice were analyzed using the Student t test.

RAG-1 and TdT genes are expressed by cells of B lineage in the bone marrow as depletion of T cells using anti-CD3 or anti-Thy-1 Abs did not significantly alter the levels of RAG-1 or TdT mRNA. TdT and RAG-1 mRNA were measured by RT-PCR using cDNA prepared from bone marrow RNA isolated from C57BL/6 mice between 1 and 19 mo of age (Fig. 1). Peak RAG-1 mRNA is observed in the bone marrow from mice 2 to 5 mo of age. Thereafter, the level of RAG-1 mRNA declines and remains, after 10 mo of age, at 10 to 20% of its peak value. In contrast, the level of bone marrow TdT mRNA does not diminish in mice between 5 and 19 mo of age and is normally expressed by nude mice (Fig. 2). We demonstrated that the level of RAG-2 mRNA is also decreased by ∼90% in old and nude compared with young normal mice in contrast to the TdT gene, which is expressed comparably in young, old, and normal mice.

A clue to the cellular basis for the decline in RAG gene expression in the thymic-deprived state was suggested by the different effect of age and the nude mutation on the expression of the RAG and TdT genes. Because both pro-B and pre-B cells express the RAG genes while the TdT gene is expressed by ∼50% of pro-B but not pre-B cells (15, 16), a decreased number of pre-B cells would lead to the changes in the levels of RAG and TdT mRNA observed. This prediction was confirmed when the numbers of pre-B and pro-B cells in the bone marrow of young, old, and nude mice were compared (Table I). There was a significant (p < 0.01) decrease in the number of pre-B cells from old and nude mice compared with young mice, while there was no significant difference in the number of pro-B cells from young, old, or nude mice.

A causal relationship between the thymic-deprived state and the decreased levels of RAG mRNA in nude and old mice was suggested by the capacity of activated T cells or ATCS to increase the expression of the RAG-1 gene by B cell precursors in nude mice (5). Figure 3 shows that injection of ATCS also increases RAG gene expression in bone marrow B cell precursors from old and nude mice.

ATCS increased the percentage of pre-B cells with detectable intracellular RAG protein (Fig. 4). Ten and 13.6% of bone marrow pre-B cells from two CTCS-treated nude mice had detectable intracellular RAG-2 protein. In contrast, 66.3 and 68.8% of pre-B cells from two ATCS-treated nude mice had detectable intracellular RAG-2 protein; this same percentage of RAG-2-expressing pre-B cells was seen in young control mice. ATCS also increased the percentage of pre-B cells from two 18-month-old mice to 89 and 91% compared with the 16% of pre-B cells from a CTCS-treated 18-month-old mouse that expressed RAG-2 protein.

To determine which T cell subset regulates the expression of RAG-1 mRNA, the capacities of supernatant medium from activated, unfractionated, purified CD4⁺ or CD8⁺ T cell cultures to induce RAG mRNA in bone marrow B cell precursors were compared (Fig. 5). Injection of ATCS from cultures of activated CD8⁺ T cells induced a >10-fold increase in RAG-1 mRNA in bone marrow B cell precursors from nude mice than did ATCS from cultures of activated unfractionated T cells. In contrast, ATCS from cultured CD4⁺ T cells induced little or no increase in RAG-1 mRNA in bone marrow B cell precursors.

To confirm the conclusion that CD8⁺ T cells are the major source of the factor in ATCS that induced RAG mRNA in bone marrow B cell precursors, we compared the level of bone marrow RAG-1 mRNA by quantitative RT-PCR in mice deficient in either CD4⁺ or CD8⁺ T cells with that in control mice (Fig. 6). The level of bone marrow RAG-1 mRNA was decreased by an average of 90% in CD8⁺ T cell-deficient (TAP1 −/−) compared with normal or CD4⁺ T cell-deficient mice (MHC class II −/−). Thus, in vivo, CD8⁺ T cells are also necessary for the full expression of bone marrow RAG genes.

IL-16 is a low m.w. lymphokine, secreted primarily by activated CD8⁺ T cells, and is thus in ATCS (17). For this reason, we measured the effect of injecting rIL-16 into nude and old mice on the levels of RAG-1 mRNA and on the number of pre-B cells in nude and old mice. In Figure 7, representative gels show that one to three injections of 100 ng of rIL-16 into four nude (one shown) or three old mice, respectively, increased their RAG-1 gene expression by bone marrow B cell precursors compared with untreated nude or old control mice. Furthermore, injection of IL-16 increased the number of pre-B cells an average of 2.6-fold in the four nude mice and by 1.6-fold in three old mice compared with untreated controls.

The key findings in this study are that the levels of bone marrow RAG mRNA and protein are decreased in thymic-deprived nude and old mice compared with control mice and that the injection of ATCS or rIL-16 into these thymic-deprived mice reverses these defects. Thus, it appears that T cells influence B cell development, a conclusion that differs from the more conventional view that B cell development is independent of T cell influence. Peripheral B cells appear to be normal in thymic-deprived nude (18) and old mice (7) because these animals have the same or greater number of peripheral B cells as do young control mice and also because the number of splenic B cells 2 mo after irradiation and bone marrow transplantation is comparable in nude and normal mice (3). However, these and other studies do not distinguish between the steady state number of peripheral B cells and the rate of their generation by the bone marrow. In contrast, nude/xid mice exhibit decreased levels of pre-B cells which could be corrected by thymus transplants from normal, young mice, suggesting a role for T cells in B cell development (19). Additionally, some studies have provided evidence of altered B cell development in old mice, primarily at the pro-B stage of development and involving the interactions of pro-B cells with stroma and IL-7 (20, 21, 22). However, this defect(s) was not related to altered T cell function. We have reported a defect in the rate of splenic B cell reconstitution from the bone marrow of thymic-deprived mice (7). Specifically, the number of donor splenic B cells in irradiated nude mice given Ly-5 disparate bone marrow cells during the first month after bone marrow reconstitution was significantly less than in similarly treated control young mice. In contrast, as previously reported, the number of splenic B cells 2 mo after reconstitution was comparable in nude and control mice (7). Other links between T cell function and the development of B cells from bone marrow precursors include the similar kinetics of thymic involution and decrease in RAG gene expression by bone marrow B cell precursors with age, the low expression of RAG genes in nude mice, and the capacity of factor(s) secreted by T cells to increase RAG gene expression and the number of pre-B cells in thymic-deprived mice.

The decreased levels of bone marrow RAG mRNA in thymic-deprived nude and old mice compared with young control mice are associated with a decreased number of bone marrow pre-B cells. Because pre-B cells make up the largest number of bone marrow cells that express RAG mRNA, it is not surprising that the level of RAG mRNA in the bone marrow would reflect the number of bone marrow pre-B cells. The association between pre-B cell number and RAG mRNA is further supported by the capacity of ATCS or rIL-16 to increase the levels of both bone marrow RAG mRNA and pre-B cells in thymic-deprived mice. However, the reduction of bone marrow RAG mRNA in thymic-deprived mice is greater than the reduction in the number of pre-B cells. This appears to be due to the fact that not only are the total number of bone marrow pre-B cells reduced in thymic-deprived compared with young control mice but also the percentage of pre-B cells that express RAG protein in thymic-deprived mice is only one-third that of pre-B cells from control mice. Pre-B cells from thymic-deprived mice with detectable RAG protein appear to have fluorescence intensity comparable to that seen for young control pre-B cells. However, fluorescent microscopy is not precise enough to rule out minor age-associated changes in RAG-2 protein expression.

It has not been determined whether the decrease in pre-B cells totally explains the decreased expression of the RAG genes in thymic-deprived mice. A low number of pre-B cells in thymic-deprived might explain the lower level of RAG mRNA in bone marrow of thymic-deprived mice. In this view ATCS or rIL-16 acts as a growth or differentiation factor to increase the steady state number of pre-B cells in thymic-deprived mice. It is known that IL-16 acts on cells that express CD4 (23). In preliminary studies, we have confirmed previous results that demonstrated the presence of CD4 on early bone marrow B cell precursors (24). This raises the possibility that IL-16 may act directly on B cell precursors to favor their differentiation. Compatible with this view is our recent preliminary observation that the level of IL-16 produced by PHA-activated T cells from old mice was 10% or less of the level seen in young mice, as determined by bioattractant activity and ELISA (W. Cruikshank et al., unpublished studies).

Alternatively, the decreased expression of the RAG genes by bone marrow pro-B cells in thymic-deprived mice might impair the rearrangement of Ig heavy chain gene segments and thereby inhibit the transition of pro-B to pre-B cells or their survival. The generation of pre-B cells requires the expression of both RAG genes and the rearrangement of the Ig heavy chain gene segments. Mice deficient in one of the two RAG gene products do not rearrange their Ig genes and do not generate pre-B cells in the bone marrow. However, the introduction of a rearranged Ig transgene into RAG knockout mice was associated with the appearance of pre-B cells (2). Thus, it is also possible that the primary defect of B cell development in thymic-deprived mice is the impaired expression of RAG genes and the consequent impairment in Ig gene rearrangement. This process might, in turn, impair the generation or survival of bone marrow pre-B cells. According to this view, injection of ATCS or rIL-16 into thymic-deprived mice increases their expression of RAG genes, favors their rearrangement of Ig gene segments, and results in their having more bone marrow pre-B cells. Studies are under way to distinguish these alternatives.

The recognition that thymic-derived factors might increase lymphocyte generation may have clinical applications. Inadequate lymphocyte reconstitution is an important source of morbidity and mortality in patients undergoing intensive chemotherapy or bone marrow transplantation. Recently, it has been reported that age is an important determinant of the rate of lymphocyte regeneration following intensive chemotherapy in patients with neoplastic disease (25). In these studies, the rate of regeneration of peripheral CD4⁺ T cells was inversely proportional to the age of the patient but directly related to the size of the thymus gland determined by radiographic imaging. Furthermore, with increasing age there was an increase in the fraction of regenerating T lymphocytes that were derived from thymic-independent pathways including peripheral renewal. It is possible that the administration of T cell factors, including IL-16, may increase the rate of T as well as B lymphocyte regeneration in older, thymic-deprived patients and thereby speed their recovery following chemotherapy or bone marrow transplantation.

We thank Dr. S. Desiderio for generously providing Abs to the RAG proteins and Drs. H. M. van Santen, M. van Roon, and A. Berns for the TAP−/− mice. We also thank Drs. Janko Nikolic-Zugic, David Posnett, and Eugenia Spanopoulou for critical reading of the manuscript and helpful discussions.