Abnormal B lymphocyte development and autoimmunity in hypoxia-inducible factor 1α-deficient chimeric mice

Results

The HIF-1α-deficient chimeric mice had short limbs (data not shown) and were found to have ES cell-derived lymphocytes as defined by the Ly-9.1 allelic marker. Analysis of the gated Ly-9.1⁺ cells from Hif1α^−/− → Rag2^+/+ B6 chimeras revealed the presence of rather normal HIF-1α-deficient CD4⁺ and CD8⁺ T cell subsets in the thymus (Fig. 1 A and B), spleen (Fig. 1 C and D), lymph nodes and peripheral blood of chimeric mice (data not shown), and dramatic distortions in phenotype of B cells (Fig. 2; see also Figs. 4 and 5).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Flow cytometry analysis of Hif1a^−/− T cell development. Thymocytes (A and B) and spleen cells (C and D) from Hif1a^−/− → Rag2^+/+ normal B6 chimera (4 weeks old) were stained with anti-Ly-9.1 mAb, anti-CD4-mAb, and anti-CD8 mAb. Stained cells were gated to enable characterization of HIF-1α^+/+ blastocyst-derived Ly-9.1 negative fraction (A and C) and Hif1a^−/− ES-derived Ly-9.1 positive fraction (B and D) by studies of the expression of CD4 and CD8.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Flow cytometry analysis of Hif1a^−/− B cell development. Flow cytometry analysis reveals abnormal B-1-like lymphocytes in the peritoneal cavity of Hif1a^−/− → Rag2^−/− chimeric mice. Panels depict representative flow cytometric experiments comparing cells from age-matched Hif1a^−/− (B, F–H, L–N, Q, and R) and Hif1a^+/+ chimeric mice (A, C–E, and O) or with age-matched wild-type 129 mice (I–K and P). Cells were gated according to forward vs. side scatter, and then ES cell-derived Ly-9.1⁺ cells were gated to be further analyzed for staining with mAb to various surface markers (13–17). Single-cell suspensions of peritoneal cavity cells were stained and analyzed by flow cytometry. The assays were done in triplicates or duplicates. The results are representative outcomes from studies of three generations of independently derived chimeric mice. Peritoneal cavity cells of control wild-type (WT Hif1a^+/+ ES → RAG-2^−/−) chimera (A) and from Hif1a^−/− ES → RAG-2^−/− chimera (B) were harvested and stained with anti-Ly-9.1 mAb, anti-CD5 mAb, and anti-B220 mAb. Then, ES-derived Ly-9.1⁺ cells were analyzed for expression of both CD5 and B220 by flow cytometry. This was followed by analysis of different CD5/B220-defined cell populations for expression of such B-1 cell markers like IgM vs. IgD (C–H, O, and Q) and Mac-1 vs. CD23 (I–N, P, and R). Demonstration of typical IgM^high and IgD^low peritoneal B-1 and B-1-like cells in wild-type (WT) and HIF-1α-deficient chimeras was accomplished by gating for analysis B220^{negative-dull}/CD5^dull fraction (C and F), B220⁺/CD5⁺ fraction (D and G), and B220⁺/CD5⁻ fraction (E and H). IgM/IgD expression on total peritoneal cavity cells are depicted in O (WT chimera) and in Q (Hif1a^−/− → RAG-2^−/− chimera). Proportions of cells are indicated in each panel. The presence of typical B1 cells among B220^{negative-dull}/CD5^dull fraction of peritoneal cells from WT mouse and among B220⁺/CD5⁺ fraction of Hif1a chimeras was confirmed in studies of Mac-1 (CD11b) vs. CD23 expression on peritoneal cells collected from 129 mouse (I–K) and from Hif1a^−/− → RAG-2^−/− chimera (L–N). In this experiment, cells were analyzed for the expression levels of Mac-1 and CD23 after cells were stained with anti-CD5 mAb, anti-B220 mAb, anti-Mac-1 mAb, and anti-CD23 mAb. B220^{negative-dull}CD5^dull fraction (I and L), B220⁺/CD5⁺ fraction (J and M), and B220⁺/CD5⁻ fraction (K and N) were gated for analysis. Expression of Mac-1 and CD23 on total peritoneal cavity cells is depicted in P (WT mouse) and in R (Hif1a^−/− → RAG-2^−/− chimera). It was confirmed in control experiments that profiles of CD5/B220 expression on Ly-9.1⁺ cells and Ly-9.1⁺/Ly-9.1⁻ cells in Hif1a^−/− ES → RAG-2^−/− chimera were similar. This allowed the comparison of cells in C–R without gating for Ly-9.1 expression.

Studies of defects in bone structure are in progress, but in the first series of experiments, we focused on investigations of changes in well described surface markers of B lymphocytes (13–17). Analysis of peritoneal lymphocytes in Hif1α^−/− → Rag2^−/− chimeric mice revealed dramatic changes in expression of B1 cell-defining CD5 and B220 surface markers (Fig. 2), whereas no differences were found in levels of expression of CD5 and B220 between control Hif1α^+/+ ES → Rag2^−/− chimeric mice and normal B6 or 129 mice (data not shown). The CD5^dullB220^−/dull subset that was observed in Hif1α^+/+ chimeras (Fig. 2A) was lost, whereas accumulation of unusual CD5⁺/B220⁺ cells was detected in Hif1α^−/− chimeras (Fig. 2B). The CD5⁺ IgM^high/IgD^low-inter expression pattern of these B220⁺cells in Hif1α^−/− chimeras (Fig. 2 C–H) is similar to peritoneal B1 cells (15), but the CD5⁺ cells in Hif1α^−/− chimeras had much higher levels of B220 expression, leading us to designate these unusual cells as B-1-like. In addition, CD5⁻B220⁺ cells in the peritoneum of Hif1α^−/− chimeras had significantly increased IgM^high/IgD^low populations (from 3.41% in control to 30.22%, Fig. 2 E and H). The same was true in studies of CD5⁺B220⁺ cells, which were found to have an increased proportion of IgM^high/IgD^low cells (from 6.9% to 32.32%, Fig. 2 D and G). The increase in proportion and number of cells with B-1 phenotype (Mac-1⁺/CD23⁻) in Hif1α^−/− chimeras was further confirmed in studies of peritoneal CD5⁺B220⁺ (Fig. 2 J and M) and of CD5⁻B220⁺ B-1 cells (Fig. 2 K and N). The increase in cells with abnormal B-1 phenotype was accompanied by a corresponding decrease in proportions of other cells (e.g. Mac-1⁻/CD23⁺, from 84% to 40%) in the peritoneum of Hiflα^−/− mice (Fig. 2 K and N). Thus, whereas the total numbers of cells harvested from the peritoneum of control and HIF-1α-deficient mice are similar (data not shown), the proportion of IgM^high/IgD^low (Fig. 2 O and Q) and Mac-1⁺/CD23⁻ (Fig. 2 P and R) cells with B-1 phenotype is much larger in Hif1α^−/− chimeras.

B-1 cells were reported to play a role in autoimmunity (16), and this prompted us to investigate the possibility of autoimmunity in Hif1α^−/− chimeras with abnormal B-1-like cells. Subsequent studies of 6–8-week-old mice revealed markedly increased levels of both IgM and IgG anti-dsDNA autoantibodies (Fig. 3 C and D) in the serum of the Hif1α^−/− chimeras. Total IgM levels were also increased in the serum of Hif1α^−/− chimeras, whereas no difference in IgG levels was found (Fig. 3 A and B). This was accompanied by extensive deposits of IgM and IgG in kidneys of Hif1α^−/−, but not Hif1α^+/+, chimeras (Fig. 3 E–H). Evidence of autoimmune disease was also provided by the demonstration of proteinuria (Fig. 3I) and accumulation of RF (Fig. 3 J–L) in Hif1α^−/− chimeras. Taken together, these observations suggest that HIF-1α is important for proper regulation of B1 cell homeostasis and that HIF-1α deficiency results in autoimmunity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Demonstration of autoimmunity in Hifla^−/− >Rag-2^−/− chimeras. (A and B) Determination of concentrations of total IgM and IgG in serum of Hif1a^+/+ (WT) → Rag2^−/− and Hif1a^−/− → Rag2^−/− chimeric mice. Results of measurements of serum IgM and IgG are presented as % of serum IgM and IgG levels, respectively, in control B6 mice. Mouse Ig ELISA quantitation kits (Bethyl Laboratories) and ELISA starter accessory package (Bethyl Laboratories) were used to measure serum Ab titer according to manufacturer's instructions. (C and D) Demonstration of increased concentration of anti-dsDNA IgM and IgG auto Ab in serum of Hif1a^−/− → Rag2^−/− vs. Hif1a^+/+ → Rag2^−/− chimeric mice and normal B6 mice. Serum samples were diluted as indicated, and concentration of anti-dsDNA Abs was tested by ELISA. The titer of anti-dsDNA Ab in mouse serum was determined by using 0.001% of protamine sulfate (Sigma) coated ELISA plates. The calf thymus DNA (Sigma) and horseradish peroxidase-conjugated goat anti-mouse IgM (Caltag) and goat anti-mouse IgG (ICN) were used for detection of anti-dsDNA IgM and IgG. The anti-dsDNA Ab titers are presented in relative units as absorbance at 450 nm wavelength. (E–H) The immunohistochemical localization of IgM and IgG deposits of paraffin-embedded sections of kidney from Hif1a^−/− → Rag2^−/− chimeras (F and H) and WT controls (E and G). Fuchsia-colored staining indicates location of IgG or IgM. The visualization of IgM and IgG on paraffin-embedded sections of kidney from Hif1a^−/− → Rag2^−/− chimeras and WT → Rag2^−/− controls was performed by using anti-IgM and anti-IgG mAb and labeled streptavidin biotin reagents (Dako). (I) Increased levels of proteins in urine from Hif1a^−/− → Rag2^−/− chimeras. The protein concentration in urine samples from eight age-matched Hif1a^−/− → Rag2^−/− chimeras and control B6 mice was determined by using the Microprotein kit (Sigma). (J–L) Demonstration of increased concentration of RF in serum of Hif1a^−/− → Rag2^−/− chimeric mice as compared with normal B6 mice. The titers of rheumatoid factors in serum were determined by using goat IgG precoated ELISA plates. IgM-RF and IgG-RF were detected by using horseradish peroxidase-conjugated anti-mouse IgM and anti-mouse IgG, respectively. Serum samples were diluted as indicated and were tested (J). Serum samples diluted 100 times were used to compare levels of RF-IgM and RF-IgG in serum samples from seven Hif1a^−/− → Rag2^−/− chimeras and six B6 mice (K and L). Six- to eight-week-old mice were used in these series of experiments.

Deficiency in HIF-1α also results in age-dependent defects in the subset composition of immature and mature conventional B-lymphocytes (B-2 cells) in bone marrow as reflected by the absence of IgM^high and B220^high (immature) population after 8 weeks of age (Fig. 4 D–G) and appearance of B220^high cells after 12 weeks (Fig. 4 H–K).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Cell-autonomous effects of the HIF-1α deficiency and the characterization of age-dependent defects in B cell development in bone marrow of HIF-1α-deficient chimeras. (A–C) The presence of normal HIF-1α^+/+ B cells and their progenitors and the absence of HIF-1α^−/− B cell subsets within the bone marrow of the same chimeric animal. Bone marrow cells from Hif1a^−/− → Rag2^+/+ normal B6 chimera (4 weeks old) were stained with anti-Ly-9.1 mAb (A) and anti-CD43-mAb and anti-B220 mAb (B and C). Stained cells were gated to enable characterization of HIF-1α^+/+ blastocyst-derived Ly-9.1 negative fraction (B) and Hif1a^−/− ES-derived Ly-9.1 positive fraction (C) by studies of the expression of B cell maturation markers CD43 and B220. (D–K) Distortion and age-dependent changes in B cell subset composition in bone marrow of Hif1a^−/− → Rag2^−/− chimeras. Bone marrow cells from both WT → Rag2^−/− chimera (D, F, H, and J) and Hif1a^−/− → Rag2^−/− chimera (8 weeks old and 12 weeks old) (E, G, I, and K) were stained with anti-Ly-9.1 mAb, and Ly-9.1-positive cells were gated and analyzed for expression of CD43 vs. B220 or IgM vs. B220 (F, G, J, and K). Proportions of cells in designated areas of graphs are indicated.

Studies of B-2 cells allowed us to discriminate between cell-autonomous and noncell autonomous effects of HIF-1α deficiency [e.g., because of the defects in HIF-1α-dependent angiogenesis and organogenesis (4, 7, 8) in Hif1α^−/− chimeric mice] by analysis of conventional B-2 cells in Hif1α^−/− → Rag2^+/+ WT B6 chimeras, in which development of Hif1α^−/− B-2 cells could be compared with Hif1α^+/+ B-2 cells within the same individual chimeric mice (Fig. 4A). The normal maturation of Ly-9.1⁻ Hif1α^+/+ B-2 lymphocytes of B6 blastocyst origin (Fig. 4B) and the block of maturation of Ly-9.1⁺ Hif1α^−/− B-2 lymphocytes of ES origin (Fig. 4C) were observed in Hif1α^−/− → Rag2^+/+ B6 chimeras (4 weeks old), suggesting that effects of HIF-1α deficiency on B cell development are cell-autonomous.

Further analysis of bone marrow B cell development in Hif1α^−/− → Rag2^−/− chimeras by using the B cell differentiation marker HSA revealed a decrease from 46.36% to 8.24% among HSA^high B220⁺ B cell progenitors and a dramatic increase in immature HSA⁻ B220⁻ bone marrow cells in Hif1α^−/− → Rag2^−/− chimeras (Fig. 5A). The decrease in HSA^high CD43⁻B220⁺ B cell progenitors (Fig. 5B) and results of IgM vs. B220 and CD43 vs. B220 analysis (Fig. 4) suggest that HIF-1α may be required for the transition from large preB to small preB [fraction D by Hardy et al. (17)] to immature B (fraction E) cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Defects in subset distribution and proliferation of Hif1α^−/− bone marrow B cells in HIF-1α-deficient chimeras. Bone marrow cells harvested from Hif1a^−/− → Rag2^−/− chimeras and control 129 mice or WT → Rag2^−/− chimeras were stained with anti-Ly-9.1 mAb to identify ES cell origin B cell progenitors, which were then analyzed for the expression of B cell differentiation markers CD43, B220, and HSA as well as for the expression of nuclear proliferation marker Ki-67. (A) Expression of HSA and B220 among Ly.9.1⁺ cells from 12-week-old Hif1a^−/− → Rag2^−/− and WT → Rag2^−/− chimeras. Proportions of cells in designated areas of graphs are indicated. (B) The decrease in HSA expressing cells among CD43⁻B220⁺ preB cells. Expression levels of HSA among CD43⁻B220⁺ bone marrow cells were evaluated after gating cells, which were stained with anti-Ly 9.1 mAb, anti-CD43 mAb, anti-B220 mAb, and anti-HSA mAb. (C) Highly proliferating cells are found among CD43⁻B220⁺ HSA^high B cell progenitors in bone marrow of control Hif1a^+/+ normal 129 mice as detected by the expression of cell proliferation marker Ki-67. Bone marrow cells from normal 129 mouse were stained with anti-CD43-mAb, anti-B220 mAb, anti-HSA mAb, and anti-Ki-67-mAb. CD43⁻/B220⁺ bone marrow cells were further gated into HSA^high and HSA⁺ cell fractions, and expression levels of Ki-67 were determined in each cell fraction. (D) Decreased Ki-67 expression among CD43⁻/B220⁺ bone marrow cells from Hif1a^−/− → Rag2^−/− chimeras. Bone marrow cells from Hif1a^−/− → Rag2^−/− chimeras and WT → Rag2^−/− chimeras were stained with anti-Ly9.1 mAb, anti-CD43-mAb, anti-B220 mAb, and anti-Ki-67-mAb. Then, Ly9.1⁺ bone marrow cells were further gated into CD43⁻B220⁺ cell fractions to compare expression levels of Ki-67.

Further studies revealed that HIF-1α deficiency results in the depletion of highly proliferating cells among B cell progenitors in bone marrow, as CD43⁻ B220⁺ HSA^high cells, which are almost completely lost in Hif1α^−/− → Rag2^−/− chimeras (Fig. 5B), are the most highly proliferating subset based on expression of a nuclear cell proliferation-associated antigen Ki-67 (18) on cells from 129 mice (Fig. 5C). Indeed, expression of Ki-67 on Hif1α^−/− CD43⁻B220⁺ bone marrow cells was much less than on CD43⁻B220⁺ cells from Hif1α^+/+ WT → Rag2^−/− chimeras (Fig. 5D).

Discussion

The multiple roles of hypoxia-inducible HIF-1α in cell metabolism and functions (4–10) and the predominance of hypoxic conditions in lymphoid tissues in vivo (2) suggested that this protein could be involved in the development and regulation of cells of the immune system. The direct testing of the possible role of HIF-1α in the immune system was precluded by midgestation death of HIF-1α-deficient (Hif1α^−/−) mouse embryos (5, 8–10). The use of the RAG-2-deficient blastocyst complementation system (11) allowed to bypass embryonic lethality and revealed that HIF-lα deficiency in Rag-2^−/− chimeric mice results in abnormal peritoneal B-1 cells, autoimmunity, and abnormal maturation of B-2 lymphocytes.

The striking cell-lineage specificity of the defects attributable to HIF-1α deficiency [no changes in T-cell subsets (Fig. 1) and distortion of B-1 and B-2 cell development] suggests that it is unlikely that these defects are because of nonspecific effects of HIF-1α deficiency on cell viability. Rather, they implicate HIF-1α in the control of B cell progenitor proliferation, which must occur by different mechanisms in B-1 and B-2 as compared with T cells based on our data. Indeed, proliferation of Hif1α^−/− B cells in vivo is inhibited in Hif1α^−/− mice (Fig. 5D). HIF-1α deficiency results in the depletion of highly proliferating cells among B cell progenitors in bone marrow, as CD43⁻ B220⁺ HSA^high cells, which are dramatically lost in Hif1α^−/− → Rag2^−/− chimeras (Fig. 5B), are the most highly proliferating subset. Hif1α^−/− CD43⁻B220⁺ bone marrow cells also express much less Ki-67 when compared with CD43⁻B220⁺ cells from Hif1α^+/+ WT → Rag2^−/− chimeras (Fig. 5D). These results indicate that HIF-1α deficiency may cause defects in proliferation of immature B cells and that the age-dependent distortion of B cell development in bone marrow may reflect the rate of B cell expansion.

The lymphoid phenotype of HIF-1α deficiency shares similarities and differences with other knockout mouse models (e.g., Btk, BSAP, and PI3K) (19–21), which were shown to play key roles in the development or homeostasis of B1 and B2 cells. Functional interactions between these gene products and HIF-1α may account for the overlapping phenotypes. For example, PI3K activity has been shown to control HIF-1α expression in prostate cancer cells (22).

The data presented here add to observations suggesting the importance of B-1 cells whose origin and functions have been the subject of controversy since they were first discovered (23). Thus, whereas the role of HIF-1 in mediating physiological responses to hypoxia is well established, these results demonstrate that HIF-1 also functions in a cell lineage-specific, cell-autonomous and differentiation stage-specific manner to regulate lymphocyte development and to control autoimmunity. Observations of autoimmunity point to the need for further investigations of the role of HIF-1α in controlling cell cycle as well as apoptosis in B-lymphocytes, as apoptosis is implicated in the elimination of autoreactive cells and HIF-1α was shown to regulate apoptosis-controlling molecules (7, 8), including p53 and NIP3. These data provide genetic evidence for the role of HIF-1α in the regulation of cell proliferation in vivo in any system and are in agreement with the implied role of HIF-1α as the switch to glycolysis in proliferating lymphocytes. These data are also consistent with the demonstration of a correlation between the expression of HIF-1α and Ki-67 in human cancers and trophoblast cells (24, 25). Recent observations of the role of glucose transporter-1 gene (GLUT-1) in lymphocyte homeostasis (24) further support the importance of HIF-1α, as GLUT-1 expression is controlled by HIF-1 (4, 7). Perhaps the most striking cell-surface abnormality revealed in the analysis of peritoneal B-1 cells was the marked elevation of B220 (CD45, a receptor-associated protein tyrosine phosphatase) (Fig. 2B), which is one of the critical modulators of B cell receptor signaling and B lymphocyte development and activation (25). The reported data suggest that HIF-1α expression in B-1 lymphocytes could be required for down-regulation of CD45 (B220) expression during normal B-1 cell development. Additional studies are necessary to delineate the exact molecular pathways by which HIF-1α deficiency results in CD45 overexpression in B-1 cells and the consequences of CD45 overexpression with respect to the pathogenesis of autoimmunity.

Taken together, these observations suggest the need to analyze the possibility of autoimmunity as being the outcome of metabolic defects in immune cells and point to the HIF-1α as a promising subject of autoimmunity studies. The future directions of research are also suggested by recent demonstration of selective up-regulation of the alternatively spliced HIF-1α mRNA isoform I.1, an immediate early response gene that encodes the HIF-1α protein without the first 12 N-terminal amino acids in activated lymphocytes (26). It remains to be established whether the effects of HIF-1α deficiency observed here are because of the requirements in this “short” I.1 isoform or in the I.2 isoform, which encodes the “classical” HIF-1α.