BVL-1–like VL30 promoter sustains long-term expression in erythroid progenitor cells

Retroviral and lentiviral model systems are close to being utilized for hemoglobinopathy human gene therapy trials.1 Incorporation of cis-acting regulatory components of the globin gene, including the locus control region (LCR), initially caused aberrant RNA processing. Modifications have allowed significant levels of recombinant globin gene expression to be achieved.2-4 Retroviral vectors using the promoter from the red cell membrane protein ankyrin have shown position-independent, copy number–dependent expression of the γ-globin gene in vivo.5 However, significant obstacles still face the contemporary constructs, including silencing regions within the vectors' long terminal repeat (LTR) regions and position effect variegation (PEV).6 Removal of silencing regions within the vectors' promoter (U3), repeat (R), and primer binding site (PBS) and insertion of insulators have not entirely prevented transcriptional silencing7 and PEV,6 respectively. Optimization of specific globin vector designs is requisite to allow for therapeutically beneficial and sustainable globin gene expression.

A variation on virally derived LTR vector components is the use of retrotransposon-based VL30 U3 promoter sequences. Mouse VL30s include more than 100 different promoter variations within the endogenous genome, have limited sequence homology to murine retroviruses,8 and are activated by transformation,9,10 growth factors,11,12steroids,13 and cytokines.14 VL30 promoters have been characterized for tissue-specific expression in vitro15,16 and in vivo.17 Hematopoietic VL30 elements14 and specialized erythropoietin (epo)–responsive erythroid progenitor cell lines have been probed.20 However, these clones have not been assessed in the context other than their native configuration. Therefore, VL30 promoter elements within a heterologous construct were reintroduced into erythroid progenitor cells. These expression assays determined that the ELM 5 (BVL-1–like) VL30 clone provided sustained reporter gene expression in erythroid progenitor cells in vitro.

VL30 promoters were isolated from MEL 585S 18 and ELM-I-1 19 murine erythroleukemic cell lines. Ribonuclease protection assays (RPAs) were performed to confirm the epo inducibility of the VL30 promoter clones in the epo-responsive ELM-I-1 cell line.20 

Linearized constructs with the novel promoters, ELM 5 and MEL/ELM, and control promoter, NVL3, within the LTRs of the VL30-derived VLIBAG cassette (17) and the retroviral vector RVBAG were electroporated into the ELM cells (10 × 10⁶ to 20 × 10⁶) using a voltage of 300 V, resistance of 13 Ω, and a capacitance of 200 F. Mass cell cultures of transfected cells were grown under selective G418 growth conditions (0.8 mg/mL).

Transfected cells were stained with X-gal (X-galactopyranoside 1 mg/mL in N,N,-dimethyl formamide) along with X-gal staining solution (20 mM K₃Fe(CN)₆; 20 mM K₄Fe(CN)₆ 3H₂0; 1.5 mM MgCl₂ in phosphate-buffered saline [PBS], pH 7.4) and incubated for 20 hours at 37°C.21 

Transfected ELM cells (2.5 × 10⁵ to 4 × 10⁵/mL) were grown for 3 days in uninduced (0.55 U/mL) and epo-induced (10 U/mL) growth conditions in T75 Falcon flasks (Fisher, Pittsburgh, PA) for 3 days. Cell lysates were monitored through duplicate assays performed in triplicate. Using the β-Gal Assay Kit (Invitrogen, Carlsbad, CA), the specific activity of the cell lysates (nanomoles of o-nitrophenyl-β-D-galactopyranoside [ONPG] hydrolyzed per milligram of total protein) from each vector construct was calculated and normalized to the total protein concentration (Pierce, Rockford, IL). The uninduced VLIBAG vector's expression was defined from each time point as 1 nmol of ONPG hydrolyzed per milligram of protein in 30 minutes at 37°C.

Raw data from each assay were assessed through a paired correlated groups Student t test using the InStat program (GraphPad Software, Version 3.01). Probability (P) values that were less than .05 were considered significant.

Degenerate VL30-specific primers were used to reverse transcribe (RT) a single-stranded 3′ U3 promoter complementary DNA (cDNA)20 (Figure 1A). Polymerase chain reaction (PCR) was used to synthesize and amplify double-stranded cDNA clones from MEL and ELM-I-1 cell RNA. VL30 promoter clones were identified and classified.20 The MEL/ELM clone was conserved between both erythroid progenitor cell lines and was grouped with members of the VL30 subgroup IV.21 The other 5 promoters were categorized in subgroup III.21 Clone ELM 5 showed homology (97%) to a previously identified VL30 element, BVL-1 (GenBank no. X17124),14,22and contained a putative erythropoietin-activated Jak2/STAT5 recognition site.23 Sequences of the ELM 5 and MEL/ELM clones encoded 3 GATA-1 sites20 and 2 B10 Ras-responsive elements.12 Confirmatory ribonuclease protection of the various VL30 promoters verified the expression of the isolated VL30 promoters.20 

The VL30 promoters MEL/ELM and ELM 5 were cloned into the promoter regions of the VL30-derived expression vector, VLIBAG (GenBank no.AF062997), for electroporation into the ELM cells (Figure 1A). A fibroblast-specific VL30 promoter, NVL3,17,24 as well as the positive control, the retroviral vector RVBAG,25 were also electroporated into the ELM cells. Transfected cells were grown in selective media and initially stained with X-gal 2 weeks after transfection to confirm the constructs' reporter gene expression (Figure 1B-E). In contrast to the control NVL3 promoter, MEL/ELM and ELM 5 constructs showed insignificant reporter gene expression levels in transformed fibroblast cell lines (data not shown).

The transfected ELM cells were grown simultaneously under uninduced (0.5 U/mL epo) and epo-induced (10 U/mL epo) growth conditions. The specific activity for each of the transfected cells was monitored over 2-week intervals for 16 weeks.

The NVL3 control promoter in the VLIBAG construct showed little activity throughout the time-course assay (Figure2). The specific activity of the positive control vector, RVBAG, was greatest after 2 weeks in culture and showed minimal expression by the 12th week (Figure 2). ELM cells transfected with the novel MEL/ELM promoter showed significant β-galactosidase expression during the fourth and sixth weeks after transfection. However, this construct also expressed minimal activity by the 16th week (Figure 2). ELM 5–transfected cells showed sustained β-galactosidase activity (Figure 2), averaging 18.64 nmol of hydrolyzed ONPG in epo-induced growth conditions. Six of the 8 time points—week 4 (P = .0441), week 8 (P = .019), week 10 (P = .0086), week 12 (P = .0007), week 14 (P = .049), and week 16 (P = .037)—reflected a significant enhancement of reporter gene activity (Figure 2). Even though the epo induction was modest, 1.2- to 1.7-fold (data not shown), the ELM 5 promoter within the VL30 vector was able to direct sustained expression throughout the 16-week time course in either growth condition.

In summary, erythroid-specific VL30 promoters were isolated, confirmed for their apparent epo-responsive behavior,20 and cloned into an expression vector, VLIBAG. Reintroduction of the novel constructs into the ELM cells revealed that the ELM 5 (BVL-1–like) promoter was able to express itself consistently during the 16-week assay period. This promoter was able to direct sustainable expression levels throughout the time course assay and continued to express itself past 16 weeks in culture. The incorporation of this promoter within the flanking LTR regions of contemporary vectors may allow for sustained therapeutic levels of transgenic globin expression.

We thank James Grunkemeyer, Daniel Meehan, and Dominic Cosgrove from Boys Town National Research Hospital for providing the VLIBAG vector and technical assistance and thank Mike Hendrickson for assistance in preparing the figures for this document.