Pregnancies modulate B lymphopoiesis and myelopoiesis during murine ageing

Abstract

We recently reported that pregnancy affects age‐related changes in the distribution of lymphoid and macrophage populations in the spleen of C57Bl/6 mice. In the present study, we examined the influence of pregnancies on the generation of various developmental B‐cell subsets and granulocyte/macrophage lineage cells during murine ageing. Using flow cytometry, changes in lymphoid (mature and early B‐cell precursors: B220^high, B220^low, surface immunoglobulin M (sIgM) µ chain +/−) and myeloid (monocyte/macrophage Mac‐1/CD11b, granulocyte Gr‐1/Ly‐6G) compartments were monitored in the bone marrow of young (2 months) and 15‐ and 23‐month‐old mice including male, multiparous and virgin female mice. Pregnancies delayed the age‐related decline in murine B lymphopoiesis and maintained B‐cell reserve capacity during ageing. We also found an increased production of myeloid cells induced by pregnancies at middle (15 months) and advanced (23 months) ages. This comparative study provides new information on changes in marrow lymphopoiesis and myelopoiesis with age. Our data emphasizes that the onset, magnitude and kinetics of age‐related changes in the haematopoietic marrow are parity dependent. These changes could influence the incidence of age‐related diseases and may account for the greater longevity of females.

Introduction

The impaired ability of the haematopoietic system to respond to frequent stimulations is an important part of the ageing process.¹ Immune dysregulation in aged humans and rodents has been clearly established for both cell‐mediated and humoral responses.^2,3 These changes were related to altered thymus T‐cell maturation with ageing and/or cumulative antigen pressure throughout life which influences B‐cell functions.⁴ Ageing is associated with a decline in antibody response to most foreign antigens but also with increased levels of autoantibodies in rodents and humans.⁵ Important changes in B‐cell generation occur during ageing.⁶ The altered representation of various developmental B‐cell subsets has been reported in conflicting data concerning the age of occurrence.^7–10 In contrast, stem cells, erythroid and myeloid lineages are not greatly affected by ageing.¹¹

Before our recent reports,^12,13 all studies on murine ageing had been conducted on virgin mice and had never examined sexual hormone changes related to life events as part of immune ageing. We have previously shown that physiological influences related to sex hormone differences throughout life induce age‐related changes in immune cell populations¹² and functions.¹³ The role of sex hormones in immune reactivity is obvious both in experimental animals and humans.^14,15 During pregnancy, higher sex hormone levels influence maternal immune reactivity and profound changes affect T and B lymphopoiesis through involution of the thymus and through a decrease in B‐cell lymphopoiesis, as previously reported.^16,17 While elevated estrogen levels selectively suppressed the generation of B‐lymphocyte precursors,¹⁷ oestrogen deficiency caused either by menopause or ovariectomy stimulated B lymphopoiesis in the bone marrow.¹⁸

Our study was designed to determine whether B‐lineage and myelomonocytic lineages are influenced by physiological hormonal changes during ageing. C57Bl/6 mice including multiparous, virgin females and males were evaluated at 2, 15 and 23 months. Our results show that pregnancies delay age‐related alterations of B‐cell development by increasing the B‐cell reserve capacity, and induce an increased production of macrophage/monocyte and granulocyte cells in bone marrow throughout life.

Materials and methods

Animals

Male, virgin and multiparous female C57Bl/6JIco mice were purchased from IFFA Credo (L'Arbresle, France) at the age of 15 and 23 months. Two‐month‐old males and females were used as controls in each experiment. All mice were maintained by the producer until use, in specific pathogen‐free colonies according to the specifications of the National Institute of Ageing. Females were mated when they were between 10‐ and 25 weeks old and then separated from syngeneic males at the age of 5 months following five controlled gestations as previously described.^12,13 Three separate experiments including aged groups (males, virgin and multiparous females) and 2‐month‐old mice (males and females) as controls were carried out with a total of 18, 6 and 6 animals, respectively, at 2, 15 and 23 months of age. Mice were used 8–10 days after receipt. Mice with evidence of pathology process were discarded.

Monoclonal antibodies (mAbs)

Phycoerythrin (PE)‐conjugated anti‐mouse CD45R/B220 (RA3–6B2), fluorescein isothiocyanate (FITC)‐conjugated anti‐mouse‐Mac‐1/CD11b (M1/70) (Caltag, San Francisco, CA) and biotin‐conjugated anti‐mouse Ly‐6G/Gr‐1(RB6‐8C5) (Pharmingen (San Diego, CA) monoclonal antibodies were used as markers for B, macrophage and granulocyte lineage cells, respectively.^19–21 Biotin‐labelled goat anti‐mouse immunoglobulin M (IgM; µ‐chain‐specific) was purchased from Southern Biotechnology Associates (Birmingham, AL).

Immunofluorescence staining and analysis

Bone marrow (BM) cells were obtained by flushing out femurs from each individual mouse with Hank's balanced salt solution (Ca²⁺‐, Mg²⁺‐ and phenol‐red‐free) (Bio‐Whittaker; Wakersville, MD) supplemented with 5% heat‐inactivated fetal calf serum (HBSS–FCS). Erythrocytes were lysed with Gey's solution for 5 min at 4°. After three washes with cold HBSS–FCS, three‐ or two‐colour staining was performed on each separate sample (viability of cells was > 90%).

For three‐colour analysis, cells (1 × 10⁶) were first incubated for 30 min on ice with PE‐anti‐Ly5/B220 and FITC‐anti‐Mac‐1/CD11b mAbs. After two washes, cells were labelled with the biotinylated‐anti‐Gr‐1/Ly‐6G mAb and counterstained with a streptavidin Tri‐color stain (5 µl, for 15 min) (Caltag, San Francisco, CA).Two‐colour analysis was performed with PE–anti‐Ly5/B220 mAb and biotinylated anti‐mouse IgM. After the final wash, stained cells were resuspended in phosphate‐buffered saline solution (PBSS) supplemented with 1% FCS at 4°.

Fluorescence‐activated cell sorting (FACS) analysis was performed on a FACScan cytometer (Becton Dickinson) as previously described.²² Each sample was analysed on 30 × 10³ unfixed cells immediately after staining. Fluorescence and light scatter signals were acquired and analysed on a Hewlett Packard (HP 340) computer using LYSIS II or Paint‐a‐gate software (Becton Dickinson, Pont de Claix, France). Light scatter analysis was performed according to the size and granularity of the cells. Fluorescent data were displayed as contour or dot plots or logarithmic histograms. Two B220 antigen expressions were analysed on fluorescence histograms, based on the inflexion point of fluorescence and determined as high and low intensities of B220 antigen corresponding to different stages of maturation, as previously reported.¹⁹ In the B220 positive population, we determined mature B cell (B220^high µ⁺) and different stages of maturation of newly formed B cells (B220^low µ⁺) and B‐cell precursors (B220^low µ^–) according to a previous study.²³

Statistics

Data were analysed using the non‐parametric Mann–Whitney U‐test and anova based on the SAS statistical package (SAS Institute Inc., Cary, NC). For each parameter, age effect was analysed by comparing each age group with young adults of the same sex. Parity or gender effects were evaluated by comparing results of mice of similar age but of different gender or parous status.

Results

Changes in myeloid and lymphoid compartments of bone marrow cells during ageing

Myeloid (large size and high granularity cells) and lymphoid (small size and low granularity cells) compartments determined by light scatter analysis in R1 and R2 gates, respectively (Fig. 1a), were significantly (P ≤ 0·01) modified with ageing (Fig. 1b). These changes mainly occurred between 15 and 23 months. The percentages of lymphoid cells (R1) decreased slowly but not significantly at 15 months and continued to decline thereafter (35 ± 3% to 38 ± 3%), reaching significance (P ≤ 0·01) compared to young adults (48 ± 2%). The myeloid compartment (R2) showed inverse profile; the percentage of myeloid bone marrow cells significantly increased in the 23‐month‐old mice (52 ± 5% to 62 ± 5%) compared to the young groups (46 ± 3%). No gender effect was observed. In late life, the percentage of myeloid bone marrow cells was slightly higher in multiparous females than in virgin groups, without statistical significance.

Figure 1.

Age‐related changes in the cell size and granularity of bone‐marrow cells. Lymphoid (R1) and nonlymphoid (R2) gates were analysed for their FSC (forward scatter) and SSC (side scatter) light angle emission. Representative result obtained from a young adult (2 months) and aged multiparous female (23 months) are shown (a). Percentage of cells in the R1 and R2 gates were analysed in the bone‐marrow of multiparous virgin female and male mice (b). Results are expressed as mean values ± SEM of 2 (n = 18), 15 (n = 6) and 23 (n = 6)‐month‐old mice. Statistical analysis showed a significant age effect: *P ≤ 0·01 between young (2 month‐old male and females) and 23‐month‐old mice (all groups).

Pregnancies delayed the age‐related decline in murine B lymphopoiesis

Bone‐marrow B‐cell development was investigated by flow cytometric analysis of B220 antigen and surface IgM (sIgM) expression. Figure 2 shows representative B220‐staining histograms in lymphoid cells within the bone marrow of one mouse from each of the young (2‐month‐old) and aged (15‐ and 23‐month‐old) groups. In line with previous observations,^19,23 three populations were identified in lymphoid gates (R1) as B220^– (non‐B cells), B220^low (pro/pre‐B cells) and B220^high (mature B cells) (Fig. 2, upper left panel). Frequency distributions within lymphoid cells defined by different levels of positive B220 antigen expression in multiparous, virgin females and males are displayed in Fig. 3.

Figure 2.

Age‐related expression of representative B220 molecule in bone marrow cells of female, virgin females and males. The upper histogram shows the level of high and low B220 expression in a young male (2‐month‐old). Percentages of B220^high and B220^low expression within lymphoid (R1) gated cells are noted for one mouse from each group.

Figure 3.

Age‐related expression of B220 molecule in bone marrow of multiparous virgin females and males. The percentage of B220^high, B220^low and total B220 positive cells are expressed as mean values ± SEM of 2 (n = 18), 15 (n = 6) and 23 (n = 6)‐month‐old mice. Statistical analysis showed a significant age effect between young and aged mice for the percentage of total B220 (^low + ^high) and B220^low subpopulations (***P ≤ 0·001). Multiparous females showed a significantly lower decrease in B220^low cells (P ≤ 0·001) than other old groups.

The frequency of total B220⁺ (^high and ^low) cells significantly declined with age in virgin and parous groups (P ≤ 0·001) (Fig. 3 upper panel). The B‐cell percentages decreased at 15 months and levelled off thereafter (23 months) compared to the young adults (2 months). While the percentage of mature B (B220^high) cells was unaffected with age (Fig. 3 lower panel), the earlier B cell compartment (B220^low) was significantly decreased (P ≤ 0·001) in all three groups compared to the young groups (Fig. 3, middle panel).

At 15 months, the B220^low cell percentages decreased to a lesser extent in multiparous mice (32%) than in virgin females (16%) or males (17%). This significant parity effect (P ≤ 0·0001) was also observed in late life: the B220^low percentage was higher in parous (29%) than in virgin female and male mice at 23 months (14% and 17%, respectively). No significant gender differences were observed.

Figure 4 shows flow cytometric analysis of representative double staining of BM cells at different ages using B220 and sIgM as differentiation markers. We determined three B subpopulations in the lymphoid gate (R1) of each mice. It is evident that virtually all the B220^high are µ⁺. Frequency distributions of B‐cell precursors (B220^low µ^–) and newly formed B cells (B220^low µ⁺) in the bone marrow of multiparous, virgin females and males are displayed in Fig. 5. In the virgin groups, the percentages of B220^low µ^– and B220^low µ⁺ significantly decreased between 2 and 15 months of age (P ≤ 0·0001), without major changes thereafter. The percentage of B220^low µ^– cells declined with age to a greater extent (70%) than B220^low µ⁺ cells (40%) in the virgin groups. Again, no significant gender effect was observed. In contrast, no age or parity effects were observed in the mature B‐cell compartment (B220^high µ⁺) (data not shown).

Figure 4.

Representative flow cytometric analysis of bone marrow cells stained with biotin‐conjugated anti‐mouse IgM (µ‐chain‐specific) and PE‐conjugated anti‐B220 in multiparous, virgin females and males at various ages. Contour plot profiles of B220 versus IgM expression (in lymphoid gate) are shown. Percentages of cells within the different fluorescence windows are given in the figure.

Figure 5.

Influence of gender and pregnancies on age‐related changes in B‐cell progenitors in bone marrow. Percentages of B220^low µ⁺ and B220^low µ^– cells are calculated in bone marrow cells of multiparous females, virgin females and males. Percentages of B220^low µ⁺ or B220^low µ^– significantly decreased at 15 and 23 months (***P ≤ 0·0001) in virgin groups (male and female) compared to the2‐month‐old mice. Multiparous mice show a significant decrease in B220^low µ⁺ cells only at 23 months (†P ≤ 0·05). The levels of B220^low µ⁺ and B220^low µ^– cells were higher in multiparous mice than in virgin groups (parity effect: P ≤ 0·01).

A significant parity effect (P ≤ 0·01) was observed in the distribution of B220^low µ⁺ cells during ageing. Multiparous mice showed a higher level of B220^low µ⁺ cells (24%) than virgin females (10%) and males (12%) at 15 months (P ≤ 0·01); percentages of B220^low µ⁺ cells were similar to young controls (22 and 19%, respectively). Later on in life, multiparous females showed a slight decrease in B220^low µ⁺ cells as compared to young mice, but the level remained higher (P ≤ 0·01) than the aged virgin groups. In multiparous females, the decline in the most immature population (B220^low µ^–) appears to decrease to a lesser extent than in virgin groups without reaching significant difference.

Long‐term effect of pregnancies on age‐related changes in murine myelopoiesis

To investigate age‐related changes in the myeloid compartment, we conducted a phenotypic analysis using anti‐Mac‐1(CD11b) and anti‐Ly‐6G/Gr‐1 mAbs as monocyte/macrophage and granulocyte lineage markers of maturation and differentiation.^21,24 Percentages of Mac‐1⁺ cells significantly increased in the bone marrow of virgin female (48%) and males (52%) at 23 months (P ≤ 0·02) compared to 2‐month‐old female (39%) or males (42%) (Fig. 6, upper panel). We observed a significant (P ≤ 0·02) age‐related increase in Mac‐1⁺ cells in multiparous females (48%) which appeared earlier (at 15 months) than in virgin groups and continued to progress at 23 months of age (60%). Multiparous females showed a significant higher level of monocytic lineage cells (Mac‐1⁺) than virgin groups in late life (P ≤ 0·001).

Figure 6.

Influence of multiparity on the expression of macrophage/myeloid (Mac‐1) and granulocyte (Ly6G/Gr‐1^high) lineage markers during ageing. Phenotypic analysis was conducted in non‐lymphoid gated cells (R2) of individual mice. Statistical analysis showed a significant age effect between young and old mice of the same sex **P ≤ 0·001 and ¶P ≤ 0·02. Multiparous females showed a higher level of Mac‐1⁺ (P ≤ 0·01) and Ly6G/Gr‐1^high+ (P ≤ 0·01) cells than in other groups at 23 months.

The distribution of bone marrow cells expressing the myeloid Ly‐6G/Gr‐1 antigen was unmodified in the virgin groups during ageing whatever the level of expression (Gr‐1^high and Gr‐1^low). In contrast, the percentage of Ly‐6/Gr‐1⁺ cells significantly increased in multiparous females (P ≤ 0·001) at 23 months compared to young adults and aged virgin groups (data not shown). This change was only related to an increased level of Gr‐1^high cells (P ≤ 0·01) (Fig. 6, lower panel). The double‐labelled Mac‐1⁺ Gr‐1⁺ cells represented the precursors of myeloid lineage which transiently express Ly‐6/Gr‐1.²⁴ This population increased only in aged multiparous females at 15 months (47% ± 2·5) and 23 months (56% ± 2·3) compared to the virgin groups at the same ages (36% ± 1·9 and 45% ± 4·5), respectively, and the young mice (35% ± 2·6 and 38% ± 2,5 in females and males, respectively) (data not shown).

Discussion

This study was designed to investigate whether physiological influences related to pregnancies or gender affect age‐related changes in bone marrow cells with respect to B‐cell and myeloid cell development. Our main results show that pregnancies delay the effect of age on B lymphopoiesis and induce an up‐production of myelomonocytic and granulocytic lineage cells throughout life, while gender has no important effect.

In the present report, we studied the age‐related decline in B lymphopoiesis. In mice, B‐lineage cells are commonly identified by the expression of B220 antigen which increases during their differentiation and maturation.¹⁹ We observed here that the decline in total B220⁺ cells is initiated at middle age and does not seem to be amplified later on. These results partly confirmed previous reports on the oldest mice from other murine strains.^7–10 Our study emphasized that the decline in B‐cell production occurring at 15 months mainly affects the earliest stage of B‐cell maturation. While the level of mature B cells (B220^high) is maintained during ageing, the effect of ageing is obvious on B‐cell precursors (B220^low) in virgin groups. In young mice, immature B cells (B220^low) represents 70–80% of the B‐cell compartment. In aged virgin groups, this percentage declines to 40–50%. Our data emphasized that ageing mainly affects B‐cell precursors (B220^low µ^–) and to a lesser extent engaged B cells (B220^low µ⁺) confirming previous reports on other strains.¹⁰

In addition, we observed no significant gender differences in age‐related changes in B lymphopoiesis. In mice, physiological hormonal decline occurred between 12 and 15 months in females and later in males.²⁵ In spite of observations linking sex hormone levels and altered B‐cell development observed in young mice,^17,18,26 there is little information on whether the loss of sex steroid function observed in human causes changes in haematopoiesis. While the level of mature B cells did not decline with age, the parallel decrease in the precursor B‐cell population suggests that there is an impairment during the committed/differentiated B‐cell stage. The bone marrow seemed able to compensate for the loss of B‐cell reserves and maintain the production of new mature B cells during ageing. This would result in aberrant functions which may have a role in the increase in autoreactive cells,²⁷ alterations in the immunoglobulin repertoire²⁸ and functional abnormalities of B cells upon stimulation.²⁹

Pregnancies seem to delay the age effect on B lymphopoiesis. Multiparous mice showed an immature/mature B‐cell ratio similar to that of the young adults, represented by a higher level of B220^low cells than in virgin groups throughout life. These results suggest that pregnancies induce an increased B‐cell reserve capacity during ageing. In multiparous females, engaged B cells (B220^low µ⁺) only decline in late life (23 months) and to a lesser extent than in virgin groups. The number of precursors (B220^low µ^–) is also less affected. It has been previously reported that pregnancy induces a decrease in B‐cell lymphopoiesis.¹⁷ The lower use of B‐cell precursors during pregnancy may preserve the ability of B cells to develop later on in life. Indeed, the number of cell divisions is limited during the lifespan (Hayflick limit).³⁰ Previously, it has been described that the telomeres of peripheral blood cells decrease throughout life.³¹ The greater capacity of B cells to be engaged a long time after pregnancy may be the result of such a phenomenon. The comparison of telomere length in aged virgin and multiparous mice is probably the best way to explore such a hypothesis. It was also recently reported that the impaired release of interleukin‐7 (IL‐7) by stromal cells is involved in the the age‐related decrease in B‐lymphopoiesis.³² Are stromal cells (or secretion of growth factors) in aged female mice different between non virgin and virgin females? Further investigations are needed to elucidate the up‐regulation of B lymphopoiesis in aged multiparous mice.

In our report pregnancies are associated with a high production of macrophage cells in the bone marrow throughout life, even though this change appeared only in the oldest virgin groups. Few studies have been conducted on age‐related changes in the myeloid compartment and their results are ontroversial. It was reported that myeloid lineages were not affected¹¹ or increased slightly in late life.³³ All these studies were reported on virgin mice. Important levels of colony‐stimulating factors (CSF) are delivered during pregnancy and concern myelopoiesis.³⁴ In previous studies, we have shown increased levels of Mac‐1⁺ cells and over‐production of IL‐3, IL‐6 and granulocyte–macrophage colony‐stimulating factor (GM‐CSF) by spleen cells of multiparous mice as of 8 months.^12,13 We suggest that this increase in CSF production in multiparous mice may induce an increase in the macrophage differentiation process. Pregnancies seem to enhance myelopoiesis during ageing as observed by a highest production of myeloid precursors (double‐labelled Gr‐1⁺, Mac‐1⁺ cells) reported here in the bone marrow and the increase in granulocyte/macrophage progenitor numbers (CFU‐GM) (data not shown). This activation may also be related to the stimulatory effect of persistent trophoblastic cells. Trophoblastic cells are identified in the maternal blood in mice and women³⁵ and are maintained for a long time after delivery (Dr G. Chaouat, personal communication). These results could be linked to those of Petrequin et al.,³⁶ where macrophage responsiveness (phagocytic activity evaluated by chemiluminescence) to adjuvants was similar in middle‐aged multiparous mice and young females. The increase in macrophage production and function with ageing appears greater in non‐virgin than in virgin females. Does this parity effect account for a longer lifespan?

These findings and those previously reported lead us to suggest that parity, which temporarily depresses the immune system, delays the effect of immune ageing by increasing the haematopoietic reserve capacity. In parous animals and humans, this phenomenon may lead to better immune functions and higher resistance to infections known to decline with age. Statistics on French centenarians revealed that 80% of this population are women of whom almost 80% were parous.³⁷ Similar statistics have recently been reported in the USA.³⁸ We believe that future studies are warranted to elucidate the extent to which the age‐related changes in haematopoiesis that we observed in parous mice may also occur in humans and may partly explain their long life expectancy.