ZASC1 knockout mice exhibit an early bone marrow-specific defect in murine leukemia virus replication

Background ZASC1 is a zinc finger-containing transcription factor that was previously shown to bind to specific DNA binding sites in the Moloney murine leukemia virus (Mo-MuLV) promoter and is required for efficient viral mRNA transcription (J. Virol. 84:7473-7483, 2010). Methods To determine whether this cellular factor influences Mo-MuLV replication and viral disease pathogenesis in vivo, we generated a ZASC1 knockout mouse model and completed both early infection and long term disease pathogenesis studies. Results Mice lacking ZASC1 were born at the expected Mendelian ratio and showed no obvious physical or behavioral defects. Analysis of bone marrow samples revealed a specific increase in a common myeloid progenitor cell population in ZASC1-deficient mice, a result that is of considerable interest because osteoclasts derived from the myeloid lineage are among the first bone marrow cells infected by Mo-MuLV (J. Virol. 73: 1617-1623, 1999). Indeed, Mo-MuLV infection of neonatal mice revealed that ZASC1 is required for efficient early virus replication in the bone marrow, but not in the thymus or spleen. However, the absence of ZASC1 did not influence the timing of subsequent tumor progression or the types of tumors resulting from virus infection. Conclusions These studies have revealed that ZASC1 is important for myeloid cell differentiation in the bone marrow compartment and that this cellular factor is required for efficient Mo-MuLV replication in this tissue at an early time point post-infection.


Background
Transcription of retroviral genomes involves numerous cellular transcription factors that bind to the unique 3′ (U3) enhancer element located in the viral promoter [1][2][3][4]. These transcription factors can dictate viral celltype tropism as well as the types of viral diseases that are elicited [5][6][7]. For example, in the case of Moloney murine leukemia virus (Mo-MuLV) mutations in the core binding site primarily induce erythroleukemia instead of thymic lymphoma [7].
Previously, our groups employed a forward genetic screen that identified ZASC1 (Zinc Finger Amplified in Esophageal Squamous Carcinoma 1 or ZNF639) as a cellular transcription factor that binds three specific DNA sequences located within the U3 region of the Mo-MuLV genome [8]. ZASC1 is a 9-C 2 H 2 zinc finger protein [9] that binds and translocates alpha-N-catenin to the nucleus [10] and also interacts with the CREB binding protein (CBP) [11]. We also showed that ZASC1 promotes viral gene expression in established cultured cell lines [8]. More recently we have found that ZASC1 is also required for efficient HIV-1 gene expression by a mechanism that is associated with recruitment of a viral Tat/cellular pTEFb complex to the viral core promoter (Bruce et al. in preparation).
To determine the role of ZASC1 in retroviral replication and disease pathogenesis in vivo, we have now generated a ZASC1 −/− mouse strain. We report here that ZASC1-knockout mice exhibit an early viral replication defect in bone marrow that is associated with a specific defect in myeloid cell differentiation.

Results
Generation and characterization of ZASC1 −/− mice To address the potential role of ZASC1 in wild type (WT) Mo-MuLV replication and disease pathogenesis in vivo, we generated a ZASC1 conditional knockout mouse ( Figure 1A). ZASC1 contains 7 exons and the open reading frame is located in exons 4-7. LoxP sites were introduced into the DNA sequences flanking ZASC1 exons 4-7 so that the complete ZASC1 open reading frame was deleted following Cre-mediated recombination. The first loxP site was inserted into the intron between exons 3 and 4 using a puromycin (puro) selection cassette. A second loxP site was inserted downstream of exon 7 and the polyadenylation signal as part of a neomycin (neo) cassette to select for insertion of this site. The neo and puro cassettes, flanked by F3 and Frt sites respectively, were excised in the presence of Flp recombinase.
The targeting vector was inserted into 129/J ES cells and 294 G418-resistant colonies were selected. Embryonic stem (ES) cell clones were screened for homologous recombination by southern blot analysis detecting a 5041 bp restriction fragment that was diagnostic of homologous recombination ( Figure 1B). Positive cell clones were subsequently screened to identify those that also contained the upstream arm of the targeting vector, and the two clones with the best morphology were chosen for injection into C57bl/6 blastocysts. Chimeric males were bred for germline transmission of the ZASC1 allele and screened by coat color.
After germline transmission was achieved, mice were crossed to ß-actin-Flp + C57bl/6 in order to remove the antibiotic resistance cassettes in the targeting construct to generate ZASC1 +/fl mice. Finally, ZASC1 +/fl mice were bred to a CMV-Cre C57bl/6 to generate ZASC1 −/− mice. Future generations of ZASC1 +/− heterozygotes were bred to generate ZASC1 −/− offspring. Quantitative PCR analysis of tail genomic DNA confirmed the complete deletion of exons 4-7 in ZASC1 from the ZASC1 fl/fl animals expressing CMV-cre ( Figure 1C). RT-PCR analysis of whole blood samples demonstrated that ZASC1 expression was also completely absent in ZASC1 −/− mice ( Figure 1D). ZASC1 −/− mice were born at normal Mendelian ratios and had similar weights and lifespans as compared to their WT littermates (data not shown). Furthermore, deletion of ZASC1 did not cause any obvious physical or behavioral defects (data not shown).

ZASC1 −/− mice have normal levels of B and T lymphocytes
Because ZASC1 is expressed ubiquitously and MLV primarily infects hematopoietic cell types, we first analyzed the major hematopoietic cell populations in lymphoid organs by flow cytometry. These studies revealed no deficiency in the proportion of T-cells or B-cells in the spleen (Figure 2A and 2B) and thymus ( Figure 2C and 2D) of ZASC1 −/− mice as compared to WT mice. Moreover, T-cells derived from young ZASC1 −/− mice showed no functional defect when tested for activation in vitro (Additional file 1: Figure S1). Therefore, we conclude that ZASC1 deficiency does not significantly alter mouse splenic or thymic B or T cell populations.

ZASC1-deficient mice exhibit an altered bone marrow common myeloid progenitor cell population
Previously, myeloid cell populations in the bone marrow were shown to be among the first cells infected by Mo-MuLV [12]. Therefore, flow cytometric analysis was used to examine the progenitor cell populations in the bone marrow. These cells were identified by a lack of lineage specific markers (lin-) and expression of sca-1 and c-kit on their surface. The lin-sca+kit+ (LSK) population contained the earliest bone marrow progenitors: hematopoietic stem cells (HSCs) were identified by gating on the CD105+ and CD150+ population whereas multipotent progenitors (MPPs) were identified by gating on the CD105+ CD150-cell subset, as described in [13]. The LK compartment (lin-sca-kit+) is known to contain myeloid and erythroid progenitors (Additional file 2: Figure S2) [14]. We found that HSC and MPP populations were comparable between ZASC1 −/− and ZASC1 +/+ mice, but there was a statistically significant 1.5-fold increase (p-value = 0.012) in the heterogenous lin-ska-kit+ (LK) compartment in all ZASC1-deficient animals tested ( Figure 3A and 3B). This compartment contains common myeloid progenitor cells and downstream myeloid precursors with no long-term repopulation potential [14]. We conclude that ZASC1 deficiency specifically leads to altered common myeloid progenitor cell differentiation in the bone marrow.
Early defect in Mo-MuLV replication in the bone marrow of ZASC1 −/− mice To address the role of ZASC1 in early retrovirus infection in vivo, a preliminary time course of infection was performed in mixed background 129 J and C57bl/6 mice with ZASC1 +/+ , ZASC1 +/− and ZASC1 −/− genotypes. Neonatal mice were inoculated p2-p3 by i.p. injection and viral titers were measured in the spleen, thymus, and bone marrow using an established focal immunofluorescence assay [15] at various time points post infection (7, 10, or 13 days) (Additional file 3: Figure S3). The virus titers in the bone marrow of ZASC1 −/− mice were significantly lower than those found in ZASC1 +/− mice 10 days post infection (p-value = 0.03).
To investigate this early defect in Mo-MuLV replication in the bone marrow in more detail, the study sample size was increased and bone marrow samples were harvested at 10 days post-infection. This analysis revealed a highly statistically significant 5.1-fold (p-value = 0.025) and 4.2-fold (p-value = 0.047) decrease in MLV replication in ZASC1 −/− mice as compared to ZASC1 +/+ or ZASC1 +/− mice respectively ( Figure 4A).
To determine if this replication defect was specific for one or more cell types in the bone marrow, we measured the virus titers in T-cell, B-cell, and in non-B/T cells in samples taken from several animals that were pooled by genotype. Mature T-cells and B-cells were immunopurified using CD90.2-or B220-coated magnetic beads respectively and flow cytometry was used to confirm cell population purity and the efficiency of T/B cell depletion (data not shown). Consistent with our previous results, the complete bone marrow samples derived from ZASC1 −/− animals supported 6.5-fold less Mo-MuLV replication as compared to ZASC1 +/+ ( Figure 4B). Strikingly, we found similar differences in virus titer when we examined the bone marrow-derived T-cell, B-cell, and non-B/T cell populations ( Figure 4B). Therefore, the difference in virus infection seen between ZASC1 −/− versus ZASC1 +/+ animals at an early time point appears to apply generally to all cell types in the bone marrow. We also asked whether ZASC1 plays a role in long-term Mo-MuLV pathogenesis by infecting neonatal mice with Mo-MuLV at the p2-p3 stage and monitoring for up to 1 year for tumorigenesis. Moribund animals were analyzed for tumor type, timing, and severity. Despite the early virus replication defect in the bone marrow, we found that ZASC1 had no significant impact on disease outcome ( Figure 4C and 4D).

Discussion
In this study, we developed a gene knockout model to investigate the roles of ZASC1 in murine development and retroviral pathogenesis. This cellular gene was previously shown to be important for efficient Mo-MuLV infection in established cultured cell lines [8]. ZASC1deficient mice developed normally, had no fertility defects, and exhibited normal T and B lymphocyte numbers as well as normal levels of bone marrow hematopoietic stem cells and multipotential progenitor cells. However, ZASC1-deficiency was associated with a significant increase in a myeloid progenitor cell population (heterogenous lin-ska-kit+ or LK cells). The LK cell compartment includes the common myeloid progenitor (CMP), which ultimately gives rise to megakaryocytes, erythrocytes, monocytes, granulocytes, and osteoclasts. We also observed a specific defect in Mo-MuLV infection in the bone marrow at an early time point post-infection. This defect was seen with both lymphocyte and nonlymphocyte populations. Despite this early block to bone marrow infection, long-term Mo-MuLV-induced disease was unaffected by this cellular factor. Previous studies had shown that early infection of the bone marrow correlates with leukemogenicity of Mo-MuLV [16]. Furthermore, cells of the myeloid lineage make up the majority of cell types infected during early stage Mo-MuLV infection after i.p. inoculation [12]. Specifically, osteoclasts and osteoclast precursors are among the first cells infected and these cell types are derived from the myeloid lineage [17]. Consequently, we envision at least two distinct models to explain the effect of ZASC1-deficiency on early Mo-MuLV replication in the bone marrow. The first model predicts that the defect seen with ZASC1 −/− bone marrow myeloid cell differentiation is responsible for the reduced level of initial infection. Reduced infection of these cells could then translate into a lower level of initial virus spread to other cell types in the bone marrow, including lymphocytes. These initial effects would presumably be subsequently masked once the virus has been amplified in the other bone marrow cell populations, explaining why we observed no differences in virus titer in this compartment at later time points or in long-term disease pathogenesis.
An alternative model is that ZASC1-deficiency leads to a more general block to virus replication (presumably at the level of viral gene expression) in all of the target cells contained in the bone marrow. However, for this model to be correct such an effect would have to be   bone marrow-specific because ZASC1-deficiency was not associated with low virus titers in samples taken early post-infection from either splenic or thymic T-or B-cell populations. Future studies will be aimed at discriminating between these two different models and exploring the putative role of ZASC1 in myeloid cell differentiation.

Conclusions
We have generated a ZASC1 knockout mouse and demonstrated that this mouse model has no obvious developmental or behavioral defects. When challenged with Mo-MuLV infection, ZASC1 deficient animals exhibited an early defect in Mo-MuLV replication in the bone marrow compartment. Further analysis of this compartment revealed an increase in the LK cell population, a group of progenitor cells that give rise to myeloid cell types including osteoclasts. Since osteoclasts in the bone marrow are among the first cell types infected by Mo-MuLV, these data suggest that ZASC1 might influence the number of these target cells available for establishing Mo-MuLV infection. Alternatively, ZASC1 might directly influence Mo-MuLV transcription in the bone marrow niche. Further studies will be needed to distinguish between these two possible models.

Generation of the ZASC1 knockout mice
A ZASC1 targeting vector was cloned using a recombineering approach, described in [18]. Briefly, loxP sites were introduced in the intron between ZASC1 exons 3 and 4 and downstream of exon 7 with puromycin and neomycin resistance cassettes. The integrity of the ZASC1 targeting vector was confirmed by PCR

Genotyping
Tails collected from 14-day old pups were incubated overnight in tail lysis buffer (100 mM Tris pH 8, 5 mM EDTA, 2% SDS and 200 mM NaCl + Proteinase K) at 55°C for in house genotyping. DNA was purified by ethanol precipitation and dried pellets were resuspended in 500 μl TE. 1 μl diluted tail DNA samples were added to a 10 μl qPCR reaction mix that also included genotyping primers and FAST SYBR Green Master Mix (Applied Biosystems). Germline transmission of the ZASC1 knockout locus was detected with primer pair 5′-TCAGTGCATCAGCATACTCAAG-3′ and 5-CTG GTGGTTCTTCATCGGC-3′ and then normalized to GAPDH using primer pair 5′-ACCCAGAAGACT GTGGATGG-3′ and 5′-CACATTGGGGGTAGGAA CAC-3′. qPCR analysis was performed using the Applied Biosystems ViiA 7 instrument. Genotyping for cre-recombinase, flp-recombinase and the floxed ZASC1 allele was done through a professional genotyping service (Transnetyx Inc.).

Mo-MuLV production and titer determination
Virus was generated in extracellular supernatants from NIH 3T3-43D cells (Fan Lab, UC Irvine), filtered through a 0.45 micron filter, aliquotted and stored at −80°C. Viral titers were determined by Focal Immunofluorescence Assay (FIA) in the presence of polybrene as described previously [15]. Virus-infected cells were identified by staining with 538 antibody-producing supernatant (Fan lab, UC Irvine) [19], and an AlexaFluor 488 Goat-anti-Mouse secondary IgG (Life Technologies). Plates were scanned and colonies counted using the Fuji FLA-5100 scanner.
The viral genome plasmid pCMMP-luciferase was used to generate the Mo-MuLV reporter viruses and mZBS-Mo-MuLV reporter virus, as described previously [8]. Virus-containing supernatants were harvested 48 hr post-transfection and treated with DNase (40 U/ml) before aliquotting and freezing at −80°C.

In vivo infections
Virus was harvested from extracellular supernatants from productively infected NIH3T3-43D cells and used to infect p2-p3 pups by intraperitonealinjection in a total volume of 25 μl Mo-MuLV (8×10 5 IU/ml). For early time points, animals were sacrificed and spleen, thymus, and bone marrow were collected in order to quantify virus infection by FIA as described above. Briefly, single cell suspensions of each tissue were generated as described before and 10-fold serial dilutions of these cells were co-cultured with NIH3T3 cells until these cells reached confluency. Similar experiments were conducted with T-cells, B-cells, and T/B-depleted samples that were purified from pooled 10-day mouse bone marrow samples using the Dynabeads FlowComp Pan-T or Pan-B cell isolation kits (Life Technologies). The cells were then stained using antibodies described in [19].

Tumor diagnosis and histopathological analysis
For tumor studies, 2-3 day old pups were infected as described above and monitored until they developed disease. Tumors were diagnosed as described previously [15]. Mice were observed daily for signs of lethargy, hunched posture, and scruffy fur associated with Mo-MuLV-induced tumors. Moribund animals were sacrificed and necropsy was performed. Spleen, thymus, liver, and kidney tissues were weighed and pieces of these organs along with lymph nodes were snap frozen and fixed in either 4% paraformaldehyde in PBS (US Biological) or 10% formaldehyde.
Fixed tissues were processed by Pacific Pathology, Inc. Fixed tissue samples were embedded in paraffin and 5 μm sections were cut and studied under light microscopy after hematoxylin-eosin (H&E) staining. Additional slides were prepared for immunofluorescence analysis by rehydrating slides, performing antigen retrieval using a sodium citrate buffer, and staining with primary rabbit antibodies to CD3 (Abcam) or CD79B (Santa Cruz Biotechnology) and with a secondary anti-rabbit antibody (Life Technologies).