Erythroid-Specific Expression of LIN28A Is Sufficient for Robust Gamma-Globin Gene and Protein Expression in Adult Erythroblasts

Increasing fetal hemoglobin (HbF) levels in adult humans remains an active area in hematologic research. Here we explored erythroid-specific LIN28A expression for its effect in regulating gamma-globin gene expression and HbF levels in cultured adult erythroblasts. For this purpose, lentiviral transduction vectors were produced with LIN28A expression driven by erythroid-specific gene promoter regions of the human KLF1 or SPTA1 genes. Transgene expression of LIN28A with a linked puromycin resistance marker was restricted to the erythroid lineage as demonstrated by selective survival of erythroid colonies (greater than 95% of all colonies). Erythroblast LIN28A over-expression (LIN28A-OE) did not significantly affect proliferation or inhibit differentiation. Greater than 70% suppression of total let-7 microRNA levels was confirmed in LIN28A-OE cells. Increases in gamma-globin mRNA and protein expression with HbF levels reaching 30–40% were achieved. These data suggest that erythroblast targeting of LIN28A expression is sufficient for increasing fetal hemoglobin expression in adult human erythroblasts.


Introduction
Development consists of a series of orchestrated stage-specific events controlled in both space and time by multiple factors including a network of heterochronic genes. Extensive research performed in the nematode C. elegans identified several factors involved in early embryonic development, including the RNA binding protein named lin-28 and its main microRNA (miRNA) target, let-7. Mutations in C. elegans lin-28 cause precocious development during larval growth, while loss of let-7 results in recapitulation of larval cell fates in adult worms [1].
The LIN28/let-7 regulatory pathway remains exquisitely well conserved throughout vertebrate evolution. The sequences of the mature let-7 miRNAs are identical in most animal species including human. During ontogeny, loss of LIN28 expression results in a concomitant increase in let-7 miRNAs in most tissues. In association with OCT4, SOX2 and NANOG, LIN28 reprograms human somatic cells to become pluripotent cells with characteristics of embryonic stem cells [2]. In embryonic and cancer stem cells, LIN28 enhances proliferation and self-renewal [3]. In contrast to stem cells, reduced expression of let-7 miRNAs in nonmalignant muscle cells and hepatocytes enhances differentiation of the cells [4,5]. Genetic manipulation of Lin28/let-7 in mice also regulates glucose metabolism [6]. Since the phenotypic effects of let-7 expression are highly dependent upon the transcriptome of the cell in which it is expressed, LIN28 is thus predicted to be functionally pleomorphic with tissue-and cell-type specificity.
The expression of human LIN28 genes has been associated with variations in body stature and timing of puberty [7][8][9][10]. The two known human homolog genes of the C. elegans lin-28 are LIN28A and LIN28B. In human CD34(+) cells, expression of LIN28A or LIN28B in culture causes increased expression of gamma-globin in conjunction with erythroid differentiation [11,12]. However, it is unclear whether LIN28 reprograms CD34(+) stem cells, or alternatively, LIN28 acts directly among committed erythroblasts to increase the expression of the gammaglobin genes. To address this topic, we explored the effects of erythroid-targeted LIN28 expression in cultured hematopoietic cells from healthy adult humans.

Ethics Statement
Approval for the research protocol and consent documents pertaining to all studies using primary erythroblasts was granted by the National Institute of Diabetes and Digestive and Kidney Diseases Institutional Review Board. Written informed consent was obtained from all research subjects prior to participation in this study.

Cell culture
Cryopreserved healthy adult human CD34(+) cells were cultured ex vivo in a 3-week serumfree system consisting of three phases: phase I from day 0 to 7; phase II from day 7 to 14; and phase III from day 14 to 21 as previously described [11].

Virus production
For lentivirus production, HEK293T cells (Thermo Scientific, Waltham, MA) were plated in 100-mm poly-l-lysine coated plates (BD Biosciences, San Jose, CA) with DMEM complete media (containing 10% FBS, l-glutamine and penicillin-streptomycin) (Life Technologies, Grand Island, NY). The plasmid mixture was prepared for co-transfection following the manufacturer's protocol for the Calcium Phosphate Transfection Kit (Life Technologies). The cotransfection mixture consists of the vector plasmid (empty vector control, KLF1-LIN28A-OE or SPTA1-LIN28A-OE vector) with packaging helper virus plasmids [15] as follows: CAG kGP1.1R, CAG4 RTR2, and CAGGS vsv-g (generously provided by Drs. Derek Persons and Arthur Nienhuis, St. Jude Children's Hospital, Memphis, TN) [16]. The day after transfection, the media was changed to DMEM with l-glutamine and penicillin-streptomycin without FBS for virus production. The lentivirus-containing supernatant was concentrated overnight following the Lenti-X Concentrator (Clontech) manufacturer's protocol and resuspended in 1/ 100 phase I culture medium of the original supernatant volume. Viral titer estimates were determined using the Lenti-X GoStix (Clontech) following the manufacturer's instructions. A MOI of 5 was calculated for the viral transductions.

Lentiviral Transduction
Cryopreserved CD34(+) cells were thawed and seeded at a concentration of 250,000 cells/ml in phase I culture medium. On day 3, 300,000 cells were resuspended at 2,000 cells/μl in phase I culture medium and transduced with viral particles. After 24 hours, the cells were resuspended in 4.0 ml phase I culture medium containing puromycin and transferred on day 7 into phase II culture medium without puromycin at 20,000 cells/ml. For each transduction, a puromycin selection control of mock-transduced cells was included until the end of phase II culture and analyzed by flow cytometry to confirm puromycin selection.

Flow Cytometry Analyses
On culture days 14 and 21, cells were stained with CD71 antibody, clone T56/14, R-PE (phycoerythrin) conjugate (Invitrogen, Carlsbad, CA) and glycophorin A (GPA) antibody, clone CLB-ery-1 fluorescein (FITC) conjugate (Invitrogen) and cell differentiation was assessed using the BD FACSAria I flow cytometer (BD Biosciences) as previously described [17]. A minimum of 5,000 live cell events was recorded and positively stained cell populations that had a fluorescence signal above two standard deviations were defined as positive.

Let-7 family of miRNAs quantitative PCR analysis
Absolute quantification for each let-7 family member was determined by constructing a standard curve prepared on the basis of the respectively synthetic targeted mature miRNA oligonucleotide of known concentration (1:10 serial dilutions, n = 6) that was run in parallel with biological samples. Each reaction was performed in triplicate. Complementary DNA and realtime PCR reaction using Taqman microRNA assay (Applied Biosystems, Grand Island, NY) were performed as previously described [20] for let-7a, let-7b, let-7c, let-7d, let-7e, let-7f-2, let-7g, let-7i and miR-98.

HPLC analysis of fetal and adult hemoglobin
Two million cultured cells at day 21 were pelleted, resuspended in distilled water and further lysed by two cycles of repeated freeze-thaw in a dry ice ethanol bath. Cell debris was removed by filtration through Ultrafree-MC devices (Millipore, Billerica, MA). Hemoglobin content was analyzed for HbF and HbA using a 20x4 mm PolyCATA column (Poly LC, Columbia, MD) fitted to a Gilson HPLC system (Gilson, Middleton, WI) as previously described [21,22]. The adult globin peak (HbA) and fetal globin peak (HbF) were quantitated and compared using Gilson Unipoint LC software (version 5.11). Total areas under the HbA and HbF peaks were used for ratio comparisons.

Statistical analysis
Replicate data are expressed as mean value ± SD with significance calculated by two-tailed Student's t test.

Erythroid LIN28A does not affect cell proliferation or terminal maturation of cultured erythroblasts
To evaluate the effects of LIN28A-OE in cell proliferation, the cell counts on culture days 14 and 21 were compared between KLF1-LIN28A-OE, SPTA1-LIN28A-OE and each respective empty vector control. No significant differences in cell proliferation were observed between the treatments when compared to control samples (Fig 3A and 3B). Erythroblast differentiation was compared between controls and LIN28A-OE cells. Flow cytometry analysis of transferrin receptor (CD71) and glycophorin A (GPA) were performed at culture day 14 (Fig 3C-3F) and day 21 (Fig 3G-3J) to determine the levels of erythroblast maturation. Interestingly, on culture day 14 of differentiation, there was a predominant population of high CD71(+) and GPA(+) cells observed in all conditions, but an accelerated maturation was observed in the KLF1-LI-N28A-OE samples as demonstrated by decreased levels of CD71 among the GPA(+) cells (compare Fig 3C-3F). On culture day 21, cell maturation was observed at comparable levels in controls, KLF1-LIN28A-OE and SPTA1-LIN28A-OE cells (compare Fig 3G-3J).

Discussion
Highly conserved across evolution, the LIN28 RNA-binding proteins are expressed in the early stages of development and are generally subjected to down-regulation during ontogeny [26]. The RNA-binding mechanism of action for LIN28 proteins is highly directed by recognition of the conserved RNA quadruplet GGAG-motif, which binds to the pri-or pre-let-7 as well as to several other RNAs throughout the cellular transcriptome [27,28]. In humans, a defined pattern of the let-7 miRNA expression during ontogeny is clearly observed in the erythroid lineage throughout the fetal-to-adult transition with significant increased expression of the let-7 miR-NAs in adult cells [20]. Additional data support the notion that the LIN28/let-7 axis is involved in fetal hemoglobin regulation as part of the developmental switching phenomenon [11].
In this study, we demonstrate that erythroid targeted over-expression of LIN28A is sufficient for robust increases in gamma-globin mRNA and HbF expression in adult human erythroid cells grown ex vivo. Lentiviral transduction vectors produced with LIN28A expression driven by the promoter region of the human erythroid KLF1 or SPTA1 genes were utilized to transduce human CD34(+) cells from adult healthy volunteers. The KLF1 gene is a transcription factor that is expressed in both primitive and definitive erythroid cell populations [29].  Functionally, KLF1 binds to several erythroid-specific gene regulatory regions, including the globin gene clusters [30,31]. The KLF1 gene promoter was chosen for this study because it is predicted to increase LIN28 expression in erythroid cells prior to the cells exhibiting high-level globin gene expression. The SPTA1 gene encodes the alpha subunit of the erythroid spectrin   protein, a major component of the red cell membrane skeleton, which is essential for the erythrocyte's biconcave disk shape and deformability [32][33][34]. SPTA1 was chosen for this study because it is exclusively found in the erythroid portion of bone marrow cells [35]. According to our colony formation assays, both KLF1 and SPTA1 promoters showed LIN28A expression with puromycin resistance almost exclusively in the cultured erythroblasts.
With both vectors, the expression of LIN28 caused increased expression of the gamma-globin gene and protein. In contrast to the reported effects of LIN28 in stem cells [2,36], we found that expression of LIN28 in erythroblasts neither caused increased growth nor inhibited maturation. The observed increase in maturation on day 14 of differentiation in the KLF1-driven LIN28 over-expression samples remains unexplained. In contrast to its role in promoting stem cell self-renewal, our data suggest that erythroblast regulation by the LIN28/let-7 pathway does not require stem cell reprogramming to increase fetal hemoglobin expression [2,36]. Studies are now being focused upon erythroid-specific features of LIN28 expression with particular interest upon identifying a mechanistic bridge between the LIN28/let-7 pathway and globin gene regulation. Our results may also be applied toward topics of globin gene therapy where erythroid-specific expression may be advantageous for safety concerns as well as therapeutic effects.