Viruses, Vol. 14, Pages 2716: The Optimized γ-Globin Lentiviral Vector GGHI-mB-3D Leads to Nearly Therapeutic HbF Levels In Vitro in CD34+ Cells from Sickle Cell Disease Patients

1. IntroductionSickle cell disease (SCD) is a monogenic disorder caused by a single amino acid substitution in the β-globin gene (glu(E)6Val(A); GAG → GTG; rs334), which results in a multi-organic disease phenotype [1,2,3]. It is characterized by the polymerization of deoxygenated sickle hemoglobin (HbS; a2βS2), leading to sickle-shaped red blood cells, and thus, to vaso-occlusion and hemolytic anemia, which are accompanied by organ damage and painful crisis [4,5]. Patients with SCD have a shortened life span, and suffer from severe clinical manifestations, such as acute painful episodes, leg ulcers [6,7], osteonecrosis, chest pain, priapism, stroke, silent cerebral infarction, systemic high blood pressure, and, less frequently, from sickle vasculopathy [8,9].The beneficial effect of fetal HbF on SCD outcome becomes apparent as early as 6 months after birth, where high levels of HbF lead to asymptomatic disease. SCD patients who continue to produce increased (>20%) γ-globin levels in adult life exhibit a less severe phenotype [4,10], primarily due to decreased HbS polymerization [8,10]. Therapies utilizing HbF inducers, such as 5-azacytidine [11,12] and hydroxyurea (HU) [13,14,15], or blood transfusion, combined with iron chelation, offer relief to SCD patients [16], as they alleviate disease symptoms.Alternative therapeutic approaches, such as gene therapy, can also be curative in SCD, as reported by Ribeil et al. in a clinical trial where a patient with severe SCD underwent gene therapy and exhibited a high proportion of anti-sickling hemoglobin post transplant, which accounted for 45% of the total hemoglobin production [17]. Gene therapy studies utilizing LVs containing either an anti-sickling βT87Q-globin or γ-globin transgenes was shown to ameliorate the SCD clinical manifestation either in vivo in SCD mouse models [18,19,20] or in vitro using CD34+ cells from SCD patients [21,22,23,24]. Additionally, the incorporation of small hairpin RNA (shRNA) in therapeutic globin LVs, either for simultaneous reduction in βS transcripts [25] or for down-regulation of the BCL11A expression [26], can lead to SCD phenotype correction. Furthermore, the role of many miRNAs that bind to transcription factors such as BCL11A, GATA-1, KLF-1, and MYB, is crucial in reactivating the γ-globin gene expression [27,28]. To this end, Sankaran et al. showed that there is a delay in silencing and persistence of fetal hemoglobin coupled with an elevation in embryonic hemoglobin in newborns, attributed to miR-15a and miR-16-1 directly targeting MYB [29,30]. Finally, following the gene editing approach, Magis et al. recently managed to successfully correct the sickle cell mutation in more than 30% of the long-term engrafting hHSCs, using a high-fidelity Cas9 nucleoprotein (RNP) and a single-stranded oligonucleotide donor [31]. Corrected erythroblasts showed a clear dominance of the corrected allele over null β-thalassemia alleles produced by non-homologous end joining (NHEJ), demonstrating a marked survival advantage in vivo [31]. We have previously demonstrated that the original LCR-free, self-inactivating (SIN), insulated γ-globin lentiviral vector GGHI [32], containing the HPFH-2 enhancer element (shown to lead to elevated HbF levels [33,34]), the -117 activating HPFH mutation of the Aγ gene promoter, and the HS-40 enhancer from the α-globin locus [35], led to in vitro correction of the thalassemic phenotype [32]. The use of the HS-40 enhancer element instead of the LCR core elements of the β-globin locus was shown to efficiently enhance the expression of the Aγ-globin gene, and was also associated with high functional titers and genomic stability [32,34], features that the globin vectors containing several LCR elements usually lack, while they are prone to genomic rearrangements and trans-activation of cancer-related genes [36,37]. Furthermore, in the novel, improved GGHI-mB-3D vector [38], we have incorporated the 3D enhancer element of the naturally occurring HPFH-1 deletion [39,40], along with the β-globin gene 3’ UTR [18], and pseudotyped it with the alternative BaEVRless envelope glycoprotein [41,42], which resulted in high and stable HbF expression at low multiplicities of infections (MOIs) in thalassemic CD34+ cells.

Based on the previous successful performance of both vectors in thalassemic CD34+ cells, in this study, we evaluated their ability to correct the SCD phenotype in vitro. To this end, CD34+ cells from non-mobilized peripheral blood of SCD patients were transduced with the GGHI or GGHI-mB-3D vectors, pseudotyped with the VSVG envelope glycoprotein. We show that transduction with the optimized GGHI-mB-3D vector leads to a significant increase in the Aγ/α ratio and HbF, along with a concomitant decrease in HbS in our patient cohorts, thus demonstrating an increased potential of improving, as well, the SCD phenotype in vitro compared to GGHI. The above effects were achieved at a transduction efficiency of 55% and a clinically relevant mean vector copy number (VCN)/cell of 1.0.

4. DiscussionIn this study, we used the previously characterized γ-globin lentiviral vector GGHI [32] and the novel optimized GGHI-mB-3D [38], both successfully assessed using thalassemic CD34+ cells [32,38], and investigated whether they can also improve or correct the SCD phenotype in vitro. We show that the optimized GGHI-mB-3D vector can significantly increase the Aγ/α ratio and HbF percentage in the SCD patient cohort, and lead to significant HbS reduction, at an average VCN of 1.0, calculated per diploid genome. This value represents the ideal target range of VCN per cell for LV-based thalassemia gene therapy [48]. Application of the Pearson’s r test across all patients showed very good correlation between normalised HbS fold decrease and normalized Aγ/α ratio (r = 0.9) or HbF fold increase (r = 0.7).Specifically, results from RP-HPLC regarding Aγ/α ratio fold increase, and following normalization to VCN (Figure S5), show that four out of eight patients (Patient 5 was not included) designated as Patients 4, 8, 9, and 12 exhibited a >2-fold increase following transduction with GGHI-mB-3D. Overall, transduction with GGHI-mB-3D led to an average Aγ/α ratio fold increase of 1.81 ± 0.97 (n = 8) following normalization to VCN. The highest increase was observed in Patient 8, who led to a corrected Aγ/α increase of 2.96-fold. Interestingly, despite the achieved maximum Aγ/α ratio fold increase in Patient 8, this was not associated with a similar increase in transduction efficiency and VCN/cell (mean transduction efficiency 90% and VCN/cell 0.9). Pearson’s r test showed poor positive correlation between Aγ/α ratio fold increase and transduction efficiency (r = 0.16) and negative correlation between the former and VCN (r = −0.002). The above results suggest that expression from GGHI-mB-3D is not entirely dependent on the vector per se, but may also be influenced by the site of integration, an observation also reported by Drakopoulou et al. [38].

Despite the significant increase in the Aγ/α ratio observed with RP-HPLC, we did not detect a reciprocally lower βS/α ratio, following transduction with GGHI-mB-3D lentiviral vector (p = 0.226, n = 9) (data not shown). This may possibly be due to the restriction of the RP-HPLC analyses to the soluble populations of globin chains. To circumvent the above findings, and in order to demonstrate a potential therapeutic effect of GGHI-mB-3D and/or of GGHI, we performed hemoglobin electrophoresis or CE-HPLC analysis of lysates from mock-transduced and transduced erythroid cultures and assessed HbS and HbF expression. As expected, GGHI-mB-3D LV led to a significantly lower HbS compared with control, demonstrating a mean fold decrease of 1.643 ± 0.88 and a 2.09 ± 2.09 mean HbF fold increase, following normalization to VCN. The respective values for GGHI were 1.890 ± 1.40 and 1.92 ± 1.59.

Following patient classification according to genotype, GGHI-mB-3D failed to demonstrate a marked in vitro improvement in the βSβS patients, suggesting higher HbF requirements for in vitro phenotypic correction in the specific patient cohort, as one would also predict from the high HbS levels before and after HU treatment in the specific patient cohort (Figure S2A,B and Table 2 and Table 3). Mean HbS fold decrease following normalization to VCN in the specific cohort was 1.46 ± 1.03 (n = 3), with the maximum value observed in Patient 4 who exhibited a 2.58-fold HbS decrease (Figure 5A, left panel), demonstrating also the maximum, i.e., 6.92-fold HbF increase (Figure 5A, right panel). On the contrary, transduction of βSβ+ cells with GGHI-mB-3D led to a 1.75 ± 0.89 (n = 5) mean HbS fold decrease and 1.57 ± 0.83 (n = 5) mean HbF fold increase, following normalization to VCN. Out of five βSβ+ patients, transduction with GGHI-mB-3D led to HbS decrease in four patients, with the highest value observed in Patient 12, who exhibited a HbS decrease of 2.84-fold (Figure 5B, left panel), also showing the highest, i.e., 2.41-fold, HbF increase (Figure 5B, right panel). Pearson’s r test showed moderate positive correlation between transduction efficiency and normalized HbS fold decrease (r = 0.56) and normalized HbF fold increase (r = 0.59).Regarding the GGHI vector in the βSβS patient cohort, the mean HbS fold decrease following normalization to VCN reached 1.74 ± 1.53, showing no statistical significance. The most profound effect was demonstrated by Patient 6, who demonstrated a 3.46-fold HbS decrease (Figure 5A, left panel), also showing the highest, i.e., 4.55-fold, HbF increase (Figure 5A, right panel). On the contrary, and in the case of βSβ+ patients, the mean HbS fold decrease and HbF fold increase were 1.98 ± 1.5 and 1.87 ± 1.4, respectively, (n = 5), both following normalization to VCN. Pearson’s r test showed no positive correlation between transduction efficiency and normalized HbF fold increase (r = −0.03) or HbS fold decrease (r = −0.09). With regards to the γ-globin mRNA levels (Figure S4), despite the marked HbF increase observed following transduction with GGHI, and, most importantly, with GGHI-mB-3D γ-globin LVs, we failed to demonstrate a marked mean increase in γ-globin transcripts relative to α-globin. The latter is in contrast with Urbinati et al. [22], in SCD cells, who showed that the γ-globin lentiviral vector V5m3-400 managed to increase γ-globin transcripts by more than 7-fold. However, this finding was based on a rather small number of three patients, while in our series we investigated a more representative cohort of eight informative patients (Patients 7 and 9 were not included due to limited cell number). We also managed to detect a small increase in at least three patients; specifically, Patients 6, 8, and 11, who showed the highest increase following transduction with GGHI, and Patients 6, 10, and 11, who showed increased γ-globin transcripts following transduction with GGHI-mB-3D LV. Most patients with increased γ-globin mRNA levels, displayed a concomitant HbF increase. A possible explanation for not detecting a similar high γ-globin increase at the mRNA level, as Urbinati et al. noted, could be attributed to the late sampling for RNA isolation. Compared to Urbinati et al. [22], we carried out RNA isolation 1 week later during erythroid differentiation, i.e., at day 20–21 instead of day 14, which may have resulted in the underestimation of expression of vector-derived γ-globin mRNA. As documented elsewhere [49,50], in vitro erythroid differentiation partially recapitulates ontogenesis and hemoglobin switching for ES-derived [49], as well as for adult peripheral blood CD34+-derived progenitors [50], with a shift to the contribution of embryonic to fetal and then to adult globin chains on the background of an overall increasing globin expression during erythroid differentiation. Since both the vector-derived Aγ gene and the endogenous one share identical cis-acting sequences, and thus undergo the same transcriptional regulation, it is conceivable that an endogenous fetal to adult switch from γ-globin to β-globin and/or βS at late stages of erythroid differentiation would result in relatively lower γ-globin mRNA levels and a higher overall βS contribution. Regarding the differences between the GGHI and GGHI-mB-3D performance, despite the significant Aγ/α ratio and HbF increase, as well as the HbS decrease demonstrated only by GGHI-mB-3D LV compared to mock-transduced cells, we did not observe significant differences between the former and GGHI, possibly due to extensive variation among patient samples, an observation also documented in a previous study by Drakopoulou et al. [38].In line with our previous observations in thalassemia patients, and of others in SCD patients [22], the mean VCN calculated for GGHI and GGHI-mB-3D lentiviral vectors was around 1.0 (p = 0.65, n = 9, unpaired two-tailed t-test), reflecting a near-ideal value for a clinical setting and a potentially therapeutic outcome, as reported by Kanter et al. [51]. VCN achieved among the βSβS and βSβ+ patient cohorts did not test significantly different between GGHI (p = 0.51, n = 4, unpaired two-tailed t-test) and GGHI-mB-3D (p = 0.66, n = 4, unpaired two-tailed t-test) LVs. With regards to transduction efficiency, GGHI and GGHI-mB-3D both achieved an overall gene transfer of 52% and 55%, respectively (p = 0.82, n = 9, unpaired two-tailed t-test). Again, no significant differences were observed in transduction efficiency among the βSβS and βSβ+ cohorts between GGHI (p = 0.75, n = 4, unpaired two-tailed t-test) and GGHI-mB-3D (p = 0.74, n = 5, unpaired two-tailed t-test) LVs.It should be emphasized that our vectors, harboring several erythroid-specific regulatory elements, to the best of our knowledge, are the only LCR-free, SIN, and insulated globin vectors that can be employed to efficiently transduce erythroid progenitor cells and successfully drive the human γ-globin gene to nearly therapeutic levels in both thalassemic [32,38] and SCD CD34+ cells, as shown in this study. In view of the ongoing clinical trials for thalassemia and SCD, one of the main strengths of these vectors is their safety feature, due to their insulation and to the lack of any LCR regulatory elements, which have been recently shown to be active in early hematopoietic progenitor cells [52], and thus are capable of trans-activating cancer-related genes, with all the known associated risks for the patient [36,37,52,53].In conclusion, despite the significant and potentially therapeutic HbF increase observed in SCD patients with the optimized γ-globin lentiviral vector GGHI-mB-3D, a major limitation of our strategy, primarily regarding the precise assessment and comparison of the HbF increase in each patient and among the βSβS and βSβ+ cohorts, is the established inherent genetic and epigenetic heterogeneity per se of these patients, coupled with the lack of information regarding important HbF genetic modifiers, such XmnI, HBS1L-MYB, and BCL11A [54], in the two cohorts. These major SNPs associated with high HbF levels, account for more than 20% of the HbF level variations among SCD patients [55], and affect both the severity and the therapeutic outcome in SCD, including patients under HU treatment [56]. In a recent systematic review by Sales et al. [57], the authors concluded that genetic variations in multiple loci, such as SNPs located at intron 2 of the BCL11A gene, can affect both baseline HbF and HbS levels in response to HU treatment in patients with SCD. Therefore, information on these SNPs, together with larger patient cohorts of different ethnic origin, would significantly contribute to a more effective assessment of the GGHI-mB-3D γ-globin vector in the context of sickle cell disease gene therapy [58].

留言 (0)

沒有登入
gif