The SLE-associated Pbx1-d isoform acts as a dominant-negative transcriptional regulator

Abstract

Pbx1 is a transcription factor involved in multiple cellular processes, including the maintenance of self-renewal of hematopoietic progenitors. We have shown that the CD4 ⁺ T-cell expression of a novel splice isoform of Pbx1, Pbx1-d, is associated with lupus susceptibility in the NZM2410 mouse and in lupus patients. The function of Pbx1 in T cells is unknown, but the splicing out of the DNA-binding domain in Pbx1-d predicts a dominant-negative function. In support of this hypothesis, we have shown that Pbx1-d transduction accelerates differentiation of MC3T3-E1 osteoblast pregenitors and mimics the effect of short hairpin RNA silencing of Pbx1. Conversely, Pbx1-d transduction reduced the expression of Sox3, a gene strongly transactivated by Pbx1, and Pbx1-d did not bind the Sox3 promoter. These results constitute a first step towards the understanding on how Pbx1-d contributes to systemic autoimmunity in the NZM2410 mouse model as well as in lupus patients.

INTRODUCTION

Sle1a1 is a lupus susceptibility locus that results in the production of activated anti-chromatin-specific CD4 ⁺ T cells with enhanced effector functions in the NZM2410 strain.^1–3 Sle1a1 expression also results in a reduced frequency of Foxp3⁺ regulatory CD4⁺ T cells,^1,4 and in vitro experiments suggested that it is at least partially due to a defective response to all trans retinoic acid during TGFβ-mediated Foxp3 induction.^3,5 The only gene located in the Sle1a1 congenic interval is the pre-B-cell leukemia homeobox 1 (Pbx1) gene, a member of the three-amino-acid loop extension family of homeodomain-containing transcription factors that modulates the DNA-binding function of Hox proteins,⁶ and interacts with Meis and Prep1, two other three-amino-acid loop extension proteins that regulate chromatin remodeling and coactivator access.⁷ Pbx1 has a central role during development and organogenesis by integrating multiple signals through its interaction with numerous partners.⁸ Therefore, Pbx1-deficient mice display an embryonic lethal phenotype.^9,10 In the immune system, Pbx1 preserves self-renewal of hematopoietic stem cells and blocks lineage-specific differentiation.¹¹ On the other hand, Pbx1-deficient embryonic stem cells fail to generate common lymphoid progenitors, resulting in the absence of B and NK cells, as well as an impaired T-cell development.¹² These seemingly opposite effects reflect the ability of Pbx–Meis complexes to contribute to either gene activation or repression by interacting with both Hox and non-Hox factors.⁸

CD4 ⁺ T-cell phenotypes associated with Sle1a1 expression are T-cell intrinsic.³ The function of Pbx1 in T cells is unknown, although Pbx1 mutants that prevent Prep1/Pbx1 DNA binding have been reported to alter T-cell thymic development.¹³ Seven protein-coding Pbx1 splice isoforms have been described, only two of which, Pbx1-a (Pbx1-002) and Pbx1-b (Pbx1-001), have known functions.^14,15 Pbx1-a and Pbx1-b show differential temporal and cell-specific expression patterns, as well as distinct and sometimes opposite functions.⁸ Pbx1-a possesses a C-terminus domain that recruits corepressor proteins SMRT and NcoR.¹⁶ It has been hypothesized that Pbx1-b lacking exon 7 that is part of the HOX-binding domain is less transcriptionally active than Pbx1-a.¹⁴

We have shown that the major splice isoform expressed in CD4⁺ T cells is Pbx1-b, which lacks exon 7 and 9 that are present in Pbx1-a.³ In addition, NZM2410 CD4 ⁺ T cells express a novel splice isoform, Pbx1-d (Pbx1-005), which lacks exons 6 and 7 that encode the DNA and HOX-binding domains, respectively.³ Pbx1 amino-acid sequence is identical between human and mouse, and we found that PBX1-d is expressed at a significantly higher frequency in the CD4 ⁺ T cells of lupus patients as compared with healthy controls.³ Pbx1-d is the only known isoform lacking the DNA-binding domain, but it can compete for cofactors with either Pbx1-a, with which it shares terminal exon 9, or with Pbx1-b, with both Pbx1-b and Pbx1-d lacking exon 7.³ This suggested that the contribution of Pbx1-d to lupus susceptibility was through its function as a dominant negative.

To test this hypothesis, we used two in vitro models in which the role of Pbx1 has been well characterized. Pbx1-b expression prevents osteoblast differentiation in MC3T3-E1 osteoblast pregenitors by nucleating the formation of a transcriptional repression complex on the promoter of osteoblast-specific genes.¹⁷ Chromatin-bound Pbx1-b complexes recruit histone deacetylases that maintain chromatin inactivation on the promoters of osteoblast-specific genes. Loss of Pbx1-b binding to these promoters allows the recruitment of histone acetylases, as well as the decreased H3K9 methylation, reflecting transcriptional activation and osteoblast differentiation. We have used this model to address Pbx1-d function in the context of transcriptional repression. Pbx1 also contributes gene activation, with myogenesis being the best characterized model.⁸ Sox3 (SRY-box 3) is a transcription factor that regulates neuronal differentiation.¹⁸ Pbx1 has been shown to be a critical factor for Sox3 basal and all trans retinoic acid-induced expression.¹⁹ We have used the Sox3 promoter to address Pbx1-d function in the context of transcriptional activation. With both models, we found that Pbx1-d expression was equivalent to knocking down Pbx1 expression, validating the hypothesis that it functions as a dominant negative. Furthermore, we showed that the Pbx1-d increased the expression of the CD44 activation marker in Jurkat T cells, and the same result was obtained by knocking down Pbx1 in this cell line. This finding constitutes a first step to determine Pbx1-d function and its contribution to autoimmunity.

RESULTS AND DISCUSSION

Pbx1-d accelerates osteoblast differentiation

To test the hypothesis that Pbx1-d acts as a dominant negative of Pbx1 transcriptional repression, we compared the effect of Pbx1-d expression with that of Pbx1 short hairpin RNA (shRNA), which has been shown to accelerate MC3T3-E1 cell differentiation into osteoblasts.¹⁷ MC3T3-E1 cells express Pbx1-b, and transduction with Pbx1-d lentivirus (LV) resulted in the expression of high levels of Pbx1-d in addition to the endogenous Pbx1-b (Figure 1a). Transduction of these cells with Pbx1-b LV resulted in over a twofold increase of Pbx1-d expression, whereas Pbx1 shRNA nearly eliminated Pbx1-b expression (Figure 1a). Osteoblast differentiation was first determined by alkaline phosphatase activity (Figures 1b and c). Pbx1-d expression significantly increased alkaline phosphatase activity as compared with empty vector (EV) transduction, and this was comparable to shRNA Pbx1 knockdown. Interestingly, both Pbx1-d and Pbx1 shRNA also increased alkaline phosphatase activity in MC3T3-E1 cells in the absence of differentiation induction. These results were confirmed by analysis of Alp message expression as well as that of Bsp and Ocn, two other osteoblast-specific genes. The expression of these three genes was significantly increased when either Pbx1-d or shRNA was expressed as compared with cells transduced with EVs (Figure 1d). Overall, these results demonstrate that Pbx1-d expression is functionally equivalent to knocking down Pbx1 expression in accelerating MC3T3-E1 cell differentiation.

Figure 1.

Pbx1-d increases osteoblast differentiation in MC3T3 cells. (a) Western blot analysis showing the expression of endogenous Pbx1-b and LV-expressed Pbx1-d (top panel with two separate transductions) in transduced MC3T3 cells, as well as Pbx1-b over- and knocked-down expression in MC3T3 cells tranfected with Pbx1-b LV or shRNA Pbx1 LV, respectively (bottom panel). (b–d) MC3T3 cells were transduced with LV vectors expressing either Pbx1-d, Pbx1 shRNA or EV and were analyzed 12 days after differentiation induction. (b) Representative image of alkaline phosphatase (ALP) activity in the absence (−) or presence (+) of β-glycerophosphate and ascorbic acid differentiation inducers. Each panel represents a confluent cell monolayer in a single well in which ALP ⁺ cells are stained purple. (c) Image analysis of Alp activity. P-values correspond to Dunnett’s multiple comparison tests. (d) Quantitative reverse transcriptase PCR analysis of Alp, Bsp and Ocn expression (n =5 independent transductions). Values were normalized to uninduced EV-transduced cells (EV −). P-values correspond to Student’s t-test. *P<0.05; **P<0.01; ***P<0.001; and ****P<0.0001.

Pbx1-d attenuates Sox3 transcription

To further explore Pbx1-d function, we compared the activity Sox3 promoter:luciferase constructs in MC3T3-E1 cells transduced with LV-Pbx1-d, shRNA Pbx1 or EV. We confirmed that the two promoter constructs that contain a Pbx1-binding site resulted in a higher luciferase activity than the construct containing only the basal promoter (Figure 2a). More interestingly, LV-Pbx1-d or shRNA Pbx1 transductions resulted in a significant lower luciferase activity of the constructs containing the Pbx1-binding site (Figures 2a and b). This showed that Pbx1-d has a lower transactivation capacity than the endogenous isoform.

Figure 2.

Pbx1-d reduced the activity of the Sox3 promoter. (a) Three Sox3 promoter:luciferase constructs, two of which (−587/ + 25 and −138/ +25) including the Pbx1-binding site, as well as the luciferase construct alone (PGL4), were transduced in MC3T3-E1 cells transduced with either LV-Pbx1-d, Pbx1 shRNA or EV. Luciferase activity was normalized to the value obtained with the −62/ +25 construct. The graph showed the individual values obtained in three independent experiments with the bars showing the mean values. (b) Pooled luciferase activity obtained with the two Pbx1-binding site constructs for each cell type. P-values correspond to Dunnett’s multiple comparison tests, **P<0.01.

Pbx1-d binding to the Sox3 promoter was investigated by electrophoretic mobility shift assay with nuclear extracts (NEs) prepared from MC3T3-E1 cells transduced with LV-Pbx1-d, shRNA Pbx1, Pbx1-b or EV. With titrated probe concentrations, NE from cells transduced with Pbx1-b resulted in consistently increased amount of probe/NE-specific complexes relative to the endogenous Pbx1-b in NE from cells transduced with LV-EV (Figures 3a and b). The small difference between LV-Pbx1-b and LV-EV suggests that near-saturating amount of endogenous Pbx1-b is bound to the Sox3 promoter. Pbx1-d expression resulted in the opposite effect with a lower amount of probe/NE complexes as compared with LV-EV or LV-Pbx1-b, with the difference reaching significance with a nonsaturating amount of probe. The amount of probe/NE-specific complexes was equivalent between M3CT3-E1 cells transduced with LV-Pbx1-d and LV-shRNA Pbx1, which was lower than the probe complexes in NE from cells transduced with LV-EV (Figure 3c). This suggested that Pbx1-d competes with endogenous Pbx1-b for cofactors, such as Meis, that favor its binding to the Pbx1/Meis-binding site. We tested this hypothesis by preincubating NEs from LV-EV- or LV-Pbx1-d-transduced cells with anti-Meis or anti-Pbx1 antibodies (Figure 3d). Anti-Pbx1 resulted in a small decrease in the amount of probe/NE-specific complexes, as previously described.¹⁹ Anti-Meis also slightly decreased the amount of complexes with endogenous Pbx1-b, but the difference was greater in LV-Pbx1-d-transduced cells. This suggests that a Pbx1/Meis complex binds the Sox3 promoter and that Pbx1-d sequesters Meis. The reduction of Meis levels by the combined presence of Pbx1-d and anti-Meis antibody greatly reduced the amount of endogenous Pbx1 bound to the probe in LV-Pbx1-d-transduced cells. These results strongly suggest that Pbx1-d does not bind to the Sox3 promoter and functions as a dominant negative decreasing the transactivation function of endogenous Pbx1-b.

Figure 3.

Pbx1-d does not bind to the Sox3 promoter in vitro. (a) Representative electrophoretic mobility shift assay (EMSA) was carried out with a probe for the Sox3 promoter containing a Pbx1-binding site with the conditions reported for each lane and NEs from MC3T3-E1 cells transduced LV-EV, Pbx1-d or Pbx1-b. The bands at the top show the specific probe/NE complex that was completely competed out by the cold probe (lane 3). The amount of hot probe was 1 × (lanes 1–5), 0.5 × (lanes 6–8) and 0.1 (lanes 9–11). The bands at the bottom show unbound free probe. (b) Band intensities of EMSA shown in (a), between the three types of NEs with graded amount of probe (mean and s.d., 3–5 samples per cell type and probe concentration). (c) NEs from cells transduced with either Pbx1-d or shRNA Pbx1 show decreased binding to the Sox3 promoter. (d) NE were preincubated with either anti-Pbx1 or anti-Meis antibody, which decreased the amount of probe/Pbx1-specific complexes. The combined lack of DNA binding of Pbx1-d and the strong binding to Meis result in a greatly diminished amount probe/Pbx1 complexes in the LV-Pbx1-d-transduced cells. For (c, d), band intensities are indicated for each lane. The P-value corresponds to Student’s t-test. *P<0.05.

PBX3, whose expression pattern largely overlaps with PBX1, has a splice isoform whose expression is favored in leukemic cells, suggesting that a truncated protein can oppose the function of the normal isoforms and lead to a pathogenic outcome.²⁰ In addition, two non-homeodomain cofactors have been shown to have critical roles in cell differentiation by preventing PBX1-HOX DNA binding. Zinc finger Pbx1 interacting protein interacts with Pbx1-b and prevents its binding to DNA in selected cell types, such as the embryonic genital tract.²¹ More directly related to the immune system, the prevention of Pbx1 binding to DNA by human Pbx1 interacting protein is necessary for hematopoietic cell differentiation.^22,23 These findings suggest that in addition to being a dominant negative, Pbx1-d may have its own regulatory function through the non-homeodomain pathways. Therefore, the association between the expression of Pbx1-d and lupus susceptibility may be mediated thorough the regulation of Pbx1/Hox/DNA-binding complexes achieved in a similar manner in other cell types by non-Pbx proteins. The contribution of this mechanism relative to a direct dominant-negative function will be directly tested in in T-cell lines as well as primary T cells.

Pbx1-d expression increases CD44 expression in Jurkat T cells

Pbx1-d affects the intrinsic phenotypes of CD4 ⁺ T cells in the NZM2410 lupus model.³ We do not know at this point as to which genes Pbx1 directly targets in T cells, making it impossible to assess the dominant-negative function of Pbx1-d in this cell type. We have previously shown that a hallmark of B6.Sle1a1 T cells was the increased expression of activation markers, including CD44,^1,2 which we confirmed here not only in the spleen but also in the gut-associated lymphoid tissue (Figure 4a). Epigenetic regulation has an important role in Cd44 expression.²⁴ Because Pbx1 transactivates gene expression through the recruitment of chromatin-remodeling factors,^7,17 we postulated that Cd44 could be one of its target gene. In addition, Cd44 expression is directly regulated by Bcl-6,²⁵ and Bcl-6 has been recently identified as a direct target of Pbx1.²⁶ Collectively, these results indicated that CD44 expression could be directly or indirectly regulated by Pbx1-d. In support of this hypothesis, CD44 expression was increased in Jurkat T cells transduced with either Pbx1-d or Pbx1 shRNA as compared with EV (Figure 4b). This suggests that Cd44 expression is directly or indirectly repressed by Pbx1 and that either Pbx1 downregulation or dominant-negative Pbx1-d expression promotes Cd44 expression.

Figure 4.

Increased CD44 expression by Sle1a1 CD4 ⁺ T cells is induced by Pbx1-d expression in Jurkat T cells. (a) Representative histograms of CD44 staining in CD4 ⁺-gated cells in the spleen, mesenteric lymph node (MLN) and Peyer’s patches (PP) of B6.Sle1a1 (empty) and B6 (gray filled) mice. (b) CD44 expression in Jurkat T cells, either transduced with EV, Pbx1-d or shRNA Pbx1 LV, normalized to CD4 expression. The numbers indicate the CD44/CD4 ratio for each sample.

MATERIALS AND METHODS

Cell culture

Newborn mouse calvarial-derived MC3T3-E1 subclone 14 preosteoblastic cells (ATCC, Manassas, VA, USA) were cultured in MEM medium (Life Technologies, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Sigma Aldrich, St Louis, MO, USA), 100 U ml ⁻¹ penicillin and 100 mg ml⁻¹ streptomycin (Gibco Life Technologies, Grand Island, NY, USA) at 37 °C with 100% humidity and 5% CO₂. The complete medium was replaced every 2–3 days and confluent cells were subcultured through trypsinisation. MC3T3 cells were allowed to grow until they reached 80–90% confluence, approximately 2 days post plating. The media was then changed to osteoblast differentiation media containing 50 μg ml⁻¹ ascorbic acid and 10 mM β-glycerophosphate (Sigma Aldrich). Differentiated cells switch from a fibroblastic to a dense cobblestone morphology. Alkaline phosphatase activity was measured 12 days later with BCIP/ NBT substrate (Sigma Aldrich) after fixation with 10% neutral buffered formalin. Quantitative reverse transcriptase PCR analysis of osteoblast-specific genes was performed under optimized conditions with SYBR Green reagents (Life Technologies). The LV-GFP-Pbx1-d, LV-GFP-Pbx1-b and LV-GFP (EV) constructs have been previously described.³ The Pbx1 shRNA LV vector msPbx1-T2 was a kind gift from Dr Jane Lian (University of Massachusetts Medical School). A total of 100 p.f.u. per cell of each recombinant LV construct were mixed with culture media containing 4 μg ml⁻¹ polybrene, then were added to 1 × 10⁶ MC3T3-E1 cells cultured to approximately 70% confluent growth. After 24 h, the transduced cells were placed into complete MEM medium. Transduction efficiency was 60–75% and resulted in stably expressed constructs. Transduction with the various constructs did not affect cell viability, growth or density. Jurkat T cells were transduced with Pbx1-d LV and Pbx1 shRNA LV as previously described.³ Reverse transcriptase PCR was performed on these cells with primers for Cd44 (F: 5′-ACTTCACCCCACAATCTTGA-3′ and R: 5′-GTGGCT TGTTGCTTTTCAGT-3′) and Cd4 (F: 5′-GAACCTGGTGGTGATGAAAG-3′ and R: 5′-CACCCCTCTGGATAAAACCT-3′).

Western blot analysis

Whole-cell lysates were prepared from MC3T3 cells 2 days after transduction with the Pbx1-d, Pbx1-b LV, Pbx1 shRNA LV or EV constructs. Western blot analysis was performed using a polyclonal anti-Pbx1 antibody (ab12001; Abcam, Cambridge, MA, USA) followed by detection with anti-rabbit horseradish peroxidase/ECL reagent as previously described.²⁷

Electrophoretic mobility shift assay

A probe was designed to span the Pbx1-binding site in the mouse Sox3 promoter by homology to the human SOX3 promoter.¹⁹ The probe was generated by annealing complementary sense (5′-ACGGCGAGCCTGTCAA TCACGAGGC-3′) and anti-sense (5′-CGTGGGCCTCGTGATTGACAGGCTC-3′) oligonucleotides. NEs from transduced MC3T3 cells were prepared 4 days after differentiation induction, with a Nuclear Extract Kit (Active Motif, Carlsbad, CA, USA) according to the manufacturer’s instructions. The binding reactions with 10 μg of NE were carried out as previously described.²⁸ In some experiments, NEs were incubated with the ab12001 anti-Pbx1 or anti-Meis (sc-10596, Santa Cruz Biotechnology, Santa Cruz, CA, USA) antibodies for 20 min at room temperature prior to the addition of the probe. In competition assays, 100-fold molar excess of unlabeled competitor was included in the binding reaction.

Luciferase assay

Three segments of the Sox3 promoter were cloned with the following primers: − 587/ +25: 5′-GAAGATCTGAAGATCTGTAGCGCACCTCCGTTATA CACT-3′ and 5′-CCCAAGCTTATGCGTTCTCTCGAGCTGGT-3′; −138/ +25: 5′-GAAGATCTGTTGATTGGCCAGGACTCAT-3′ and 5′-CCCAAGCTTATGCGTT CTCTCGAGCTGGT-3′; and −62/ + 25: 5′-GAAGATCTCAACCCTCCCGGATC TGA-3′ and 5′-CCCAAGCTTCCCAAGCTTATGCGTTCTCTCGAGCTGGT-3′. BglII and HindIII restriction sites were inserted in the upstream and downstream primers, respectively, for cloning purposes. PCR products were individually inserted into a TK-luc⁺ PGL-4 basic reporter vector (Promega, Madison, WI, USA). The accuracy of each construct was verified by sequencing. The − 587/ + 25 and −138/ + 25, but not −62/ + 25, constructs contain a strong Pbx1-binding site.¹⁹ MC3T3-E1 cells expressing Pbx1-d or Pbx1 shRNA LV were transduced with 50 ng TK-luc ⁺ vector or 1.6 μg individual Sox3 promoter constructs. Cells were lysed 48 h later to measure luciferase activity as previously described.²⁸ Normalization was performed to the TK-luc empty construct.

Flow cytometry

Flow cytometry was performed on splenocytes and cells from mesenteric lymph nodes and Peyer’s patches from 7-month-old B6.Sle1a1 and B6 mice with anti-CD4-FITC (RM4–5) and anti-CD44-PE (IM7, both from eBisocience, San Diego, CA, USA) as previously described.³

Statistical analysis

Statistical analysis was performed using Prism software (Graphpad Software, La Jolla, CA, USA). Unless specified, graphs show means and s.e. Statistical significance was reached with P-values ≤0.05 in the two-tailed tests indicated in the figure legends.