In vivo repopulation ability of genetically corrected bone marrow cells from Fanconi anemia patients

Abstract

Fanconi anemia (FA) is a rare inherited genomic instability syndrome representing one of the best examples of hematopoietic stem cell deficiency. Although FA might be an excellent candidate for bone marrow (BM) genetic correction ex vivo, knockout animal models are not sufficient to guide preclinical steps, and gene therapy attempts have proven disappointing so far. Contributing to these poor results is a characteristic and dramatic early BM-cells die-off when placed in culture. We show here that human primary FA BM cell survival can be ameliorated by using specific culture conditions that limit oxidative stress. When coupled with retrovirus-mediated transfer of the main complementation group FANCA-cDNA, we could achieve long-term reconstitution of the stem cell compartment both in vitro and in vivo. Gene-corrected BM cultures grew for >120 days, and after cultured cell transplantation into NOD∕SCID mice, clonogenic human cells carrying the FANCA transgene could be detected 6 months after transduction. By comparison, untransduced cells died in culture by 15 days. Of necessity for ethical reasons, experiments were conducted on a very limited number of primary BM cells. By using low cytokine regimen and conditions matching regulatory requirements, a contingent of gene-corrected cells slowly emerges with an unmet potential for in vivo engraftment. Future therapeutic applications of stem cells might be expanding from these data. In addition, we provide a model of gene-corrected human primary cell growth that carries the potential to better delineate the combined role of both DNA damage and oxidative stress in the pathogenesis of FA.

Fanconi anemia (FA) is a rare autosomal recessive disease that results in early bone marrow (BM) failure (1). Aplastic anemia develops at an average age of 7 years. Other features of the disease include abnormal skin pigmentation, growth retardation, skeletal malformations, and abnormalities in the kidney and heart together with an increased risk for the development of leukemia and squamous cell carcinoma (2, 3). So far, 11 distinct FA complementation groups have been identified (A–C, D1, D2, E–G, I, J, and L), and currently 9 FA genes underlying these subtypes have been cloned (4–7). FANC proteins are involved in DNA repair∕cell cycle checkpoints, redox metabolism, and differentiation processes (8, 9). Cells cultured from FA patients exhibit increased spontaneous chromosomal aberrations and hypersensitivity to DNA cross-linking agents such as mitomycin C (MMC) or diepoxybutane. This in vitro phenotype has become the basis for including FA as one of the genomic instability syndromes (10). Pancytopenia is associated with reduced BM cellularity, with a profound deficiency in the erythroid compartment as the main characteristic. The primary therapeutic option available is allogeneic BM or cord-blood transplantation, which carry significant side effects due to the disease background and as a consequence of increased sensitivity to the cytotoxic conditioning (11, 12) needed to facilitate allogeneic stem cell engraftment. Gene therapy of FA-aplastic syndrome using autologous multipotent hematopoietic progenitors would be a more suitable therapeutic option. Previous gene therapy attempts in FA-C patients have proved rather disappointing though, because neither permanent hematopoietic stem cell correction nor even sustained in vitro culture of FA blood-forming cells has been observed to date (13, 14).

Several obstacles need to be overcome for this approach to become successful including (i) the scarcity of FA primary BM cells in hypoplastic or aplastic marrow; (ii) their extraordinary fragility, and (iii) limitations of current gene transfer technology. In this study, we show that in overcoming the initial cell catastrophe caused by oxidative stress and DNA damage we were able to restore the long term ability of retrovirally transduced stem cells to proliferate and differentiate both in vitro and in vivo in a stem cell disorder where cells can otherwise not replicate and die. Of interest, these data have been obtained by using unfractionated BM cells so that autologous gene-corrected stromal∕mesenchymal cells provide natural support for the patient’s hematopoietic stem cell growth. Experiments have been performed on very small cell samples by using serum-free conditions that would match requirements to translate in the clinic once safety-improved vectors will have been engineered. We provide a model of gene-corrected human primary cell growth that carries the potential to better delineate the combined role of both DNA damage and oxidative stress in the pathogenesis of FA.

Results

Limiting Oxidative Stress Prevents Human FA Primary BM Cells Early Cell Catastrophe.

In this study, we focused mainly on controlling cell-fragility and early death on the one hand and improving gene transfer on the other hand, beginning with the biology of target cells. To avoid any unnecessary manipulation of the cells that might cause additional cell injury and loss we used unfractionated BM. Of necessity for ethical reasons relating to the disease features themselves and the paucity of the samples available from each patient, experiments were performed on a limited number of cells, i.e., 2 × 10⁵ nucleated BM cells (10–50 μl according to each patient BM cellularity). First, several modifications of cell culture conditions were compared on samples from 9 FA-A patients. Cells were incubated under low oxygen pressure (5%) and tested in parallel in the presence of three different antioxidative agents. Cell counts and clonogenicity were then evaluated (Fig. 1A). Among the three market-approved agents under test [N-acetyl cysteine (NAC), Amifostine, and Vastarel], only NAC showed statistically significant positive effects on cell viability∕survival (Fig. 1B and Table 1). After a 2-day incubation of the BM cells, no gross cell count differences were recorded (Table 1, b), with or without the use of culture conditions including either hypoxia or NAC. However, by day 8 of culture (Table 1, c), a statistically significant improvement (p = 0.045) in BM cell viability and survival was seen with the combination of hypoxia and NAC. Colony-forming unit (CFU) assays proved to be even more discriminating. When cultured by using conditions including hypoxia and NAC, nine of nine FA-A samples tested were able to generate CFU, of which eight contained erythroid colonies {CFU-E and BFU-E [BFU-E, burst-forming unit (erythroid)] in some instances} and four contained mixed colonies (CFU-GEMM) (Table 1, d and e). Without antioxidant conditions, only three of nine BM samples gave rise to a few myeloid clusters, and only one gave rise to erythroid clusters. Under no circumstances were either BFU-E or GEMM-CFUs observed when cells were cultured under “standard conditions” (p values: erythroid, 0.027; global, 0.009).

Fig. 1.

Effects of the antioxidants conditioning and FICD transduction on the clonogenicity of FA patients’ BM cells. (A Left) One and 2 small colonies obtained with Amifostine or Vastarel; with 55 large colonies with NAC (50 × 10³ cells were seeded maintained under 5% O₂; unconditioned cells not represented: no colony). (A Right) NAC conditioned BM cells differentiate into different lineages BFU-E (left) and CFU-GM (right). (B) Example of early protection of primary FA-BM cells (patient 1) 1,000 × 10³ cells were grown by using NAC plus hypoxia (▴) versus NAC alone (■) or normoxia (•). Hypoxia alone would not deliver. NAC has an intrinsic effect in combination with hypoxia. (C) FICD-mediated rescue of BM cells catastrophe. At day 8, untransduced control cells are not surviving (a) and transduced cells will settle in LTC (b), starting from a majority of affected cells of very small size with morphological abnormalities. As functional correction progresses, morphological modifications show here at day 61 of LTC: (c) larger cells in suspension; (d) smaller, refringent, and semiadherent CD34⁺ cells.

Table 1.

Hypoxia and NAC conditioning protects FA-A patients’ BM against initial cell damage

Patient no.	Cell counts						CFU assays§							Vector∥
	Day 0*		Day 2†		Day 8‡		Erythroid		Myeloid		GEMM	Total¶
	Per ml BM	Percent CD34+	Conditioning				Conditioning							FICD	Lenti
			No	Yes	No	Yes	No	Yes	No	Yes	Yes	No	Yes
1. Ale	2.36 × 106	1.34	ND	ND	27,000	208,000	0	8	0	55	0	0	63	+	−
2. Jav	7.4 × 106	0.3	Debris	140,000	Debris	72,000	0	121	0	48	3	0	172	+	+
3. Jul	1.9 × 106	ND	ND	150,000	ND	56,000	0	2	0	15	0	0	17	+	+
4. Aud	1.3 × 106	ND	ND	150,000	ND	93,000	0	1	0	17	1	0	19	+	−
5. Cha	4.9 × 106	5	288,000	308,000	14,000	147,000	0	14	0	37	0	0	51	+	+
6. Lau	0.7 × 106	0.9	54,000	66,000	ND	ND	0	2	5	13	0	5	15	+	+
7. Cas**	2 × 106	1	1,206,000	1,562,000	80,000	116,000	0	1	5	13	1	5	15	+	−
8. Rho**	1.9 × 106	10.2	468,000	374,000	4,000	36,000	0	0	1 cluster	12 cluster	0	1	12	+	+
9. OCT**	0.4 × 106	0.5	1,188,000	1,518,000	6,800	124,000	6	5	11	26	1	17	32	+	−
p value					0.045		0.027					0.009

*BM sample cell concentration and percentage of CD34⁺.

^†BM cells counts starting from 2 × 10⁵ cells and expressed per ml, with or without conditioning.

^‡The effects of conditioning on cell survival appear to be statistically significant (p = 0.045) at day 8.

^§CFU assays (from 50 × 10³ at most) performed with or without NAC and hypoxia conditioning. Erythroid colonies grew from eight conditioned patients’ samples and only from one unconditioned (9) (p = 0.027). Myeloid colonies readout is monitored in all conditioned samples and mix (CFU-GEMM) colonies grow from these only (2, 4, 7, and 9) because neither BFU-E nor GEMM-CFU can be observed out of unconditioned samples.

^¶The overall p value (0.009) is highly significant (Wilcoxon paired test).

^∥Indication of which patients’ samples have been exposed to both FICD and lentivirus-based vectors in the transduction experiments.

**Cultures from patients 7, 8, and 9 were initiated from 1 × 10⁶ cells.

Gene Transfer Results in Functional Restoration of FA Patient’s Primary BM Cells.

By using this procedure, FA patients’ primary BM cells could be maintained in culture long enough to match the minimal conditions needed for gene transfer. We chose to focus on FA-A patients, the most common FA complementation group (>70%). A bicistronic retrovirus vector was designed, FOCHA-FANCA-IRES-CD24 (FOCHA-FICD) (illustrated in Fig. 4, which is published as supporting information on the PNAS web site). This original vector was derived from the FB29-Fr-MuLV strain as a high-titer (10⁶ cfu∕ml) GALV producer clone (15).

FICD-mediated functional correction of the FA-A defect was first demonstrated on genotyped primary fibroblasts from FA-A patients. Ninety percent of these cells could be readily transduced by using this vector. Unlike untransduced control FA-A fibroblasts that developed massive apoptosis, FOCHA-FICD-treated cells resisted to both chronic exposure (10 nM) and 48 h of high-dose (100 nM) MMC. As a confirmation, transduction of two FANCA-null lymphoblast cell lines (HSC72 and DAF 1001) resulted in their acquisition of resistance to MMC induced cell cycle G₂ blockade, whereas untransduced controls were blocked 48 h after culture in 100 mM MMC. In addition chromosome breakages induced by chronic MMC treatment (10 nM) were abolished (<4% versus 100%) as determined by cytogenetic analysis performed after 6 weeks of chronic MMC treatment. The presence of the FANCA protein in the transduced cells was demonstrated with specific antibodies (kindly provided by Maureen Hoatlin, Oregon Health and Science University, Portland). Furthermore, a protein truncation test was developed and confirmed the overall integrity of the FANCA sequences generated after gene transfer (data not shown).

The next step in our project targeted the functional restoration of patients’ primary BM cells. The transduction procedure and subsequent maintenance of cells in long-term culture (LTC) were optimized in a series of 30 experiments. We used NAC-hypoxia-cultured BM samples from 10 different patients (2 × 10⁵ cells per patient), among which 9 were of A (FA-A) complementation group and 1 was from complementation group C as control. Within 15 days of culture, untransduced FA-A cells and the FA-C cells disappeared from culture dishes (Fig. 1C). By contrast, FOCHA-FICD-transduced hematopoietic cells from all FA-A patients were able to develop into LTCs. A layer of autologous stroma spontaneously developed since unfractionated BM was used to establish the cultures (Fig. 1C). These LTCs were kept under minimal cytokine regimen and analyzed weekly. With most samples, LTCs could be maintained for over 5 weeks. Interestingly, we found that cultures established from transduced cells could be (i) returned to culture under normal oxygen concentration after 4 weeks (but not earlier) and (ii) exposed to low-dose MMC over 15 days (from day 28 to 43) without noticeable impairment of either morphology, growth rate, or cell counts as compared with untreated controls.

Dramatic improvement of the following parameters was noted when monitored weekly throughout LTCs: cell number, diameter, morphology, viability, clonogenicity, and membrane resistance (based on tolerance to both centrifugation and slide∕coverslide pressure). Evidence for a selective growth advantage of the cells treated by FICD transduction was the observation that a subset of cells began to regenerate the culture beginning around days 20–30 (Fig. 2A). The percentage of cells entering apoptosis was followed by FACS analysis. Before day 50 of the culture, Annexin V- and propidium iodide-negative cells represented one-third of the cells growing in suspension, whereas from day 60 on, double-negative cells reached 83.2% and even 87.6% at day 70 (Fig. 2B, where both suspension and semiadherent cells were analyzed; see below). It is of interest to note that the FACS analysis also had to be adjusted to account for dramatic changes in cell diameters from 5–7 μm in starting material to 10–15 μm at day 70; accordingly, the gain was changed from 30 to 10.

Fig. 2.

Long-term inhibition of apoptosis in transduced primary BM cells. (A) Viability of cells in LTC: cells counts over time (60 days) in transduced (■) and untransduced (•) cells. Untransduced cells cannot grow and rapidly disappear around day 10; a subset of transduced cells begins to regenerate the culture from days 20–30 LTC. (B) FACS analysis based on propidium iodide and Annexin V labeling: >80% of transduced cells are alive, e.g., days 63 and 70 of LTC, whereas only debris are detected in untransduced controls. (C) Analysis of CD24 expression by confocal microscopy at day 65: a positive signal is present in all cells alive as a marker of FICD transduction (nuclei colored in red).

Confocal microscopy demonstrated the presence of CD24 in all cells, reflecting the presence of the cDNA sequences introduced by the gene transfer vector (Fig. 2C). In addition, semiquantitative PCR assays run on both bulk cells growing in suspension and individual CFU-derived colonies seeded at day 21 revealed the presence of the FOCHA-FICD-specific sequences (data not shown). BM samples from patients initially included 65–75% of lymphoid cells. As genetically modified cells grow in culture, the lymphoid lineage disappears and myeloid cells develop as monitored by FACS using antibodies against CD3, CD14, CD15, and glycophorin A (plus CD36 and CD71 in some cases; data not shown). Progressive increase in erythroid progenitors up to 30–50% over 28 days accounts for a striking improvement because the absence of erythroid progenitors is the primary hallmark of FA. Among the nine FA-A patients that we tested, similar data have been observed in cultured BM samples from eight (with the exception of patient 8 from Table 1) and are clearly the result of functional complementation resulting from successful introduction of the FANCA transgene.

Long-Term Reconstitution of the Stem Cell Compartment in Vitro.

Based on these encouraging data, a confirmatory experiment was conducted with the BM sample providing the best LTC-IC potential to both (i) recover a higher number of cells at late stages and (ii) evaluate the hematopoietic potential of this unpertubated LTC. The BM of choice corresponds to patient 5 from which no CFU could be obtained without antioxidative conditions (Table 1), and untransduced controls disappeared from culture dishes by day 15. Cytokine concentrations were intentionally kept low to maintain cells in the most undifferentiated state. This particular experiment was maintained until day 120 of LTC in transduced wells only. Data are shown in Fig. 3, except those for untransduced controls because all cells were dead and no CFU were obtained.

Fig. 3.

FICD-mediated reconstitution of a hematopoietic stem∕progenitor cells compartment (day 60). (A) Cytological analysis (cytospin and May–Grünwald–Giemsa coloration) of cells growing in suspension: (Left) a single cell displays an apoptotic nucleus (arrow); (Right) a cell in mitotic prophase at this late stage. (B) The other subset of refringent, smaller semiadherent cells growing on the autologous stroma. At day 70, these cells were analyzed by confocal microscopy confirming CD34 positivity and a high nucleocytoplasmic ratio (C). (D) FACS showing 7% CD34⁺ cells (relative to the whole cell population, whether growing in suspension or adherent). (E) FACS analysis at day 110. CD38 expression was found negative in 1.6% of the sample; a CD34⁺∕CD38⁻ stem∕undifferentiated progenitor compartment has been rescued. (F) Day-110 cells gave rise to small clusters dividing in methylcellulose.

Again, the LTC developed on autologous stroma. At day 60 when first evaluated, hematopoietic cells could clearly be subdivided into two subsets. (i) The first contained a majority of larger cells (12–15 μm) with a low nucleus∕cytoplasm ratio growing in suspension (Fig. 3A Left). These progenitor cells undergo differentiation and proliferation to a limited extent because of the minimal cytokine regimen under use. The presence of cells in mitotic prophase confirmed the dividing activity of transduced cells and no chromosome break at this late stage of the LTC (Fig. 3A Right). Day-75 suspension cells continued to grow and gave rise to small myeloid colonies in methylcellulose (6 of 1,000 cells). (ii) The second contained smaller cells that attached to the stroma as round, refringent, and semiadherent cells (Fig. 3B, LTC). These morphologic and growth characteristics are compatible with the presence of less differentiated hematopoietic cells. Indeed, when analyzed by confocal microscopy at day 70 of the LTC, these cells had a much higher nucleus∕cytoplasm ratio and tested positive for CD34 expression as shown on Fig. 3C. Again, all cells observed were also positive for CD24. FACS analysis revealed 7% CD34⁺ cells (Fig. 3D) at day 70; analysis performed at day 110 confirmed the presence of 7% CD34⁺ and 1.6% CD34⁺∕CD38⁻ cells (Fig. 3E). Day-110 cells gave rise to small clusters developing and dividing (doublets, as shown in Fig. 3F) in methylcellulose.

To further evaluate the hematopoietic potential of FICD-transduced cells, we used a separate in vitro assay (16) in which the cells were grown on a MS-5 xenogenic mouse stroma after limiting dilution. Transduced BM cells were first grown during 4 weeks in liquid LTC, a timeline that we formerly identified as appropriate for functional recovery. At that time, 100 × 10³ transduced cells were divided into two subsets. One was serially diluted and seeded on the MS-5 presettled feeder layer, and after 5 weeks each well was transferred to methylcellulose. Three large CFU-G colonies were obtained from the last 10 × 10³ cells (>300 cells each). The second subset was directly seeded in methylcellulose. Interestingly, the readout was of two gigantic GEMM colonies, three G colonies, and one GM colony. Mixed colonies kept developing for 14 days. A secondary CFU assay produced a total of 109 colonies: 2 mix, 24 G, 20 GM, and 63 small colonies counted at day 70 posttransduction. Semiquantitative PCR confirmed the presence of the FOCHA-FICD sequences in each colony picked at random.

These data indicate that hematopoietic cells with myeloid lineage potential have emerged from the initial unfractionated BM sample. These in vitro assays indicate that a contingent of less differentiated cells might carry potential for hematopoietic reconstitution, a potential that requires in vivo challenge to be conclusively demonstrated. The selective growth advantage observed in vitro was specific for cells subjected to retrovirus-mediated gene transfer. We thus chose to expand transduced cells in LTC, then to transplant them into NOD∕SCID mice as a secondary assay.

Long-Term Engraftment of Stem Cells in Vivo.

To evaluate this potential, BM cells were taken from three FA-A patients at distinct phases of the disease: patient 5, early diagnosis; patient 9, initial phase of leukemic transformation and patient 7, terminal aplastic anemia (Table 1). Again, after a preliminary growth period under antioxidant culture conditions, 2 × 10⁵ unfractionated BM cells were transduced and grown for 9 weeks in LTC. Before the injection into 12-week irradiated NOD∕SCID mice, CD34⁺ cells represented 4 × 10³ cells (20% of total) for patient 5, 66 × 10³ cells (25%) for patient 9, and 32 × 10³ cells (18%) for patient 7, compared with 5%, 0.5%, and 1%, respectively, at day 0 (Table 2). Cells from patient 5 were injected into a single mouse, and cells from patients 9 and 7 were injected into two mice each. After 12 weeks, just before they were killed, all mice looked healthy. All hematopoietic organs appeared grossly normal, in particular livers and spleens were sound, normally sized, homogeneous, and supple. Cells were recovered from BM and isolated from spleen and liver to perform independent FACS analyses, human cytokine-based CFU assays, and PCR. No sign of leukemia was detected, and in particular no leukemic cells developed from the cells from patient 9.

Table 2.

FICD-transduced cells perform in NOD∕SCID mice

Patient no.	Human BM FICD-transduced cells				Mice used	NOD∕SCID mice assay
	Day 0		Day 63			BM				Spleen colonies	Liver colonies
	Cell count	CD34+ %	Cell count	CD34+ %		Colonies	CFU-readout
							Erythroid	Myeloid	GEMM
5. Cha	2 × 105	5	2 × 104	20	1	185	3 BFU-E, 4 CFU-E	176	2	71 (21 clusters)	0
7. Cas	2 × 105	1	3.05 × 105	18	2	0	—	—	—	0	0
9. OCT	2 × 105	0.50	3.2 × 105	25	2	13	—	6 GM, 4 M, 3 G	—	0	0

hCFU assay from NOD∕SCID mice BM samples, 14 weeks (over 3 months) after infusion of transduced cells primarily grown in LTC for 9 weeks; CD34⁺ percentages were measured before infusion. Human CFUs were obtained only with cells from the early diagnosed patient.

Human MHC⁺∕CD45⁺ cells (2%) were detected by FACS in the BM from mouse 5 only. The presence of human hematopoietic CFU in mouse tissues was analyzed under culture conditions that are permissive to human cell survival only. In mouse infused with cells from patient 5, 185 colonies, among which 3 BFU-E, 4 CFU-E, and 2 CFU-GEMM were produced from 200 × 10³ BM cells; 71 colonies (and 21 clusters) were generated from the spleen and none from the liver; in mice corresponding to patient 9, 13 colonies were counted from the BM only (6 GM, 4 M, and 3 G), and in mice corresponding to patient 7, the readout was null (Fig. 4A). To verify the presence of the gene transfer vector, PCR was performed on the 13 largest colonies picked individually from mouse 5 BM, and all tested positive (see Fig. 5, which is published as supporting information on the PNAS web site). Bulk cells recovered from the methylcellulose were also PCR-positive for the FANCA vector. In contrast, samples from patients 9 and 7 did not engraft significantly in NOD∕SCID mice, although the number of cells initially infused was much higher, among which 20% on average tested positive for CD34⁺. Interestingly, this is correlated with the absence of CFU from these patients in the assays performed on MS-5.

Discussion

These data demonstrate, despite the limited numbers of cells available for analysis, that gene-corrected primary BM cells from FA patients selectively grow in culture and sustain hematopoietic potential that translates into a positive engraftment in vivo in NOD∕SCID mice. By comparison, untransduced cells died in culture by 15 days. Of interest, these data have been obtained by using unfractionated BM cells so that autologous gene-corrected stromal∕mesenchymal cells provide natural support for the patient’s hematopoietic stem cell growth. Experiments have been performed on very small cell samples by using serum-free conditions that would match requirements to translate in the clinic once safety-improved vectors will have been engineered. In this study, we show that in overcoming the initial cell catastrophe caused by oxidative stress and DNA damage, we were able to restore the long term ability of retrovirally transduced stem cells to proliferate and differentiate both in vitro and in vivo in a stem cell disorder where cells can otherwise not replicate and die.

A speculative model of how FA-associated proteins repair interstrand crosslinks during DNA replication had been recently strengthened by the identification of two new FA genes (17). We could monitor full recovery of cell growth and proliferation with each patient’s sample from day 28 posttransduction. At that time, cells were brought back to normoxia and addition of NAC was stopped. Based on these observations, gene-corrected cells are capable of both proliferation and differentiation that may account for an improved capacity for DNA to replicate. Another striking empiric observation is the delayed functional recovery of cell-membrane resistance and cell diameter. Prior studies have hypothesized the role of abnormal redox status that may account for FA-related membrane fragility (18–20). Interestingly, our data show that cell-membrane fragility would not recover before day 60 of culture in vitro, the time by which cell diameter would also normalize, as per FACS analysis. Thus, cells might recover from the ability to sustain oxidative stress at a much later stage where DNA can supposedly already replicate. Analyzing the sequential phases of human primary stem cells recovery might provide insights into the molecular mechanisms involved in FA: this unique model carries potential to better delineate the combined role of both DNA-repair deficiency and oxidative stress in the pathogenesis of FA.

It is noticeable that comparative side-to-side experiments performed with two lentivectors transmitting the FANCA-cDNA resulted in failure of hematopoietic cells to further grow between days 30 and 50 of LTC (see Fig. 6, which is published as supporting information on the PNAS web site). This observation was unexpected because faster functional recovery was initially observed as compared with FICD vector. Lentivirus-mediated transduction was performed for 12 h only and FICD for three cycles over 3 consecutive days (see Materials and Methods for details). These data were identical whichever lentivector construct was used: (i) the reference pSIN-18 PGK W-PRE (kindly provided by Luigi Naldini, H. San Raffaele Scientific Institute, Milan) or (ii) an original construct (LENP29 without W-PRE) using Fr-MuLV U3 sequences as strong internal enhancer∕promoter substituting the housekeeping PGK sequences in pSIN-18.

Interestingly, other groups have performed experiments in knockout Fanca^−/− or Fancc^−/− mice that report efficient gene transfer into as-yet-unaffected hematopoietic stem cells by using various vectors such as MLV [MSCV (21)], lentivirus (22), retrotransposons,** and, more recently, Foamy virus-based defective vectors.^†† Hematopoiesis in these mutant mice is virtually normal under usual conditions and only becomes impaired under conditions of special stress. Therefore, the ex vivo gene therapy protocols in mutant mice are not strictly comparable with the conditions that exist in the clinical setting where BM cells are already disabled at the time of diagnosis. Perhaps a model involving the prior administration of a 3-month low-dose MMC regimen to BM donor mice before gene transfer and reinfusion into syngeneic recipients would represent an experimental model more comparable to the clinical situation.

All experiments performed here have been conducted with an extremely limited amount of patient’s primary BM cells. Despite this limitation, our data clearly demonstrate that BM cells from the majority of patients with FANCA can be successfully grown and will differentiate along myeloid lineages under special conditions that reduce unnecessary physical manipulation, provide for antioxidant growth conditions, secure natural environment in using autologous stromal support, and restore gene function via retrovirus-mediated gene transfer. In addition, our procedures meet criteria matching regulatory bodies most stringent requirements with the use of unfractionated autologous BM cells, serum-free media, market-approved antioxidative agents, and minimal cytokine regimen.

The clinical challenge remains because a sufficient number of patient’s primary cells would need to be efficiently transduced and carry a hematopoietic potential able to (i) reconstitute a human BM and (ii) sustain a quantitatively sufficient and qualitatively appropriate production of peripheral blood cells. Furthermore, our data provide strong evidence that this potential varies widely from one patient to the next. In view of the severe adverse events observed in the Paris-X-SCIDs trial (23, 24), safety-improved gene transfer vectors need to be developed. We believe that it would not be ethical to perform further experiments on patients’ BM at this stage. Clinical prospects might be considered at a later stage, making use of those safety-improved vectors that we are currently developing. Finally, clinical success will be possible only in those patients diagnosed early enough in the course of the disease, where a hematopoietic stem cell pool might be spared and appropriate for gene correction, as demonstrated by our in vivo data.

Materials and Methods

Recruitment of Patients: BM Samples.

Experiments were performed on nine FA-A patients (25) and one FA-C as a control. All patients tested were children unable to sign informed consent, which was obtained from their parents. BM samples (1–5 ml) were taken and stored in liquid nitrogen as aliquots of Ficoll-separated cells. No other selection was performed. Our protocol is based on BM samples only. In fact, collection of peripheral blood mobilized progenitor cells in FA-A patients could not be considered for the following reasons, including ethical: (i) poor yield in FA children, (ii) potential risk related to the use growth factors such as G-CSF (burst of preleukemic cells), and (iii) the need for a central venous catheter under general anaesthetic.

Antioxidant-Hypoxia Cell Conditioning Procedure.

The market-approved antioxidants used were NAC, Amifostine {2-[(3-aminopropyl)amino]ethanethiol dihydrogen phosphate}, and Vastarel [(2,3,4-trimethoxybenzyl)piperazin dihydrochloride].

Cells were thawed in stem cell growth medium (CellGenix) highly defined serum-free medium in the presence of the following cytokine mixture (TPO, 5 ng∕ml; LIF, 10 units∕ml; SCF∕MGDF, 25 ng∕ml; Flt-3, 5 ng∕ml; IL-3, 5 units∕ml; G-CSF, 10 units∕ml; and Epo, 0.5 units∕ml) and antioxidant (1 mM NAC, or 0.5 mM Amifostine, or 50 ng∕ml Vastarel), and grown under hypoxia (5% O₂) for 1 week before tested in clonogenic assays.

Statistical Analysis.

Comparisons were performed by using the Wilcoxon paired test.

Vectors.

The bicistronic construct, FOCHA-FANCA-IRES-CD24 (FOCHA-FICD), is based on the FOCH29 (Fr-MuLV FB29-strain) described in ref. 15. It includes a tailored FANCA cDNA (kindly provided by Hans Joenje, VU University Medical Center, Amsterdam), the EMCV-IRES (from pCITE), and the CD24 surface antigen encoding cDNA (kindly forwarded by Keith Humphries, Terry Fox Laboratory, BC Cancer Research Centre, Vancouver) (for details, see Fig. 4 and Supporting Materials and Methods, which are published as supporting information on the PNAS web site).

Retrovirus-Producer Clones and Primary Evaluation of Transduction Capacity.

The third generation of TE-FLY-GA GALV producer cell lines [engineered by F-L Cosset (26)] were engineered with FANCA-encoding vectors via primary transfection. A stable high titer (10⁶ cfu∕ml) FICD producer clone was selected (clone 3B5) through several rounds based on numbers of copy transmitted (PCR), sustained titers, and genomic stability (over years) (see Supporting Materials and Methods for details).

Cell Cultures.

Lymphoblastoids (EBV) HSC72 and DAF1001 were grown in RPMI with 20% FCS (HyClone), 2 mM l-glutamine, and 50 μg∕ml gentamycine. Patients’ primary fibroblasts were grown in Iscove’s modified Dulbecco’s medium with 10% myoclone serum, 10% horse serum, 2 mM l-glutamine, and β FGF (1 μl∕ml). Producer cells and TE-671 cells (used for titers-assays) were grown in DMEM with 10% NBS (HyClone) serum, 2 mM l-glutamine, and 50 μg∕ml gentamycine. 3B5-producers medium was replaced by half NBS-DMEM and SCGM when cells reached 80% confluence and further by 100% SCGM. Fresh supernatant was harvested and filtered through 0.45-μm filters. Transduction was performed with protamine-sulfate (5 μg per ml) at a multiplicity of infection (MOI) of 5–10, and 90% of cells were transduced by using one round of exposure to undiluted virus soup. Each lot of virus was checked for titers and absence of helper.

Hematopoietic Cell Transduction.

Cells were thawed in SCGM with 1 mM NAC and 100 ng∕ml TPO, and seeded overnight in 12-well plates. Anti-TGF-β (10 μg∕ml) was added ≈16 h before transduction. Three cycles were applied over 3 consecutive days with FICD particles at a MOI of 5–10. The viral supernatant was directly harvested in SCGM medium containing 1 mM NAC, 25 ng∕ml TPO, 25 units∕ml LIF, 25 ng∕ml SCF∕MGDF, and 25 ng∕ml Flt-3. After transduction, cells were placed in LTC. Comparative experiments have been performed simultaneously with our two FANCA lentivectors (p24 over 2,500 ng∕ml). Identical conditions were used before transduction. Cells were exposed to lentivectors (MOI of 50–80) overnight. Accordingly, cells were seeded in LTC 2 days before FICD-transduced ones.

LTC Conditions: Liquid LTC.

Cultures were performed with NAC (1 mM) under hypoxia (5% O₂) during the first 28 days. At this point, addition of NAC was stopped and cultures were brought back to normoxia. Cells were grown in SCGM under low concentrations of growth factors: 5 ng∕ml TPO, 10 units∕ml LIF, 25 ng∕ml SCF∕MGDF, 5 ng∕ml Flt-3, 5 units∕ml IL-3, 10 units∕ml G-CSF, and 0.5 units∕ml Epo. Half of the supernatant was replaced each week. Weekly analysis consisted of counts, phenotyping (myeloid markers: FACS), CFU assays, apoptosis (Annexin V and propidium iodide), transduction, and expression; PCR and RT-PCR on FANCA ORF, CD24, and vector sequences; presence of CD24 (FACS, confocal); and cytology (see Supporting Materials and Methods for details).

LTC on xenogenic stroma (of murine origin MS-5) was used to support the growth of limiting dilutions of cells (27, 28). After transduction and 4 weeks in liquid LTC, cells were seeded on MS-5 monolayer adhering to a film of 1% gelatin.

Clonogenic Assays.

CFU assays were performed in MethoCult H4436 (StemCell Technologies, Vancouver), addressing early progenitors. One to 5 × 10⁴ cells were seeded (35 mm). Clonogenic assays were performed immediately after transduction and every 15 days except for wells untouched until day 60 and those for in vivo infusion. Colonies were analyzed both at day 14 and day 22 to ascertain sustained growth.

Human Cell Transplantation in Immunodeficient Mice.

Six-week-old C.B.17 scid∕scid (SCID)∕NOD mice bred in our facility were used. Mice were maintained in sterile isolators and fed with sterile food. All manipulations were performed under a laminar flow hood. Sentinel mice were screened regularly for the absence of subclinical viral and bacterial infections. Before transplantation, mice were irradiated at 3.5 Gy by using a Cs¹³⁷ irradiator. Cells were injected i.v. through the retroorbital plexus in 0.2 ml of PBS. The same route was used to harvest blood for chimerism analysis. Mice were eventually killed for analysis by cervical dislocation.

Real-Time PCR.

To study the presence of the FICD gene transfected in cells, we developed a real-time semiquantitative PCR. DNA was isolated from cells by using the Pico Pure DNA extraction according to the manufacturer’s instructions (Arcturus, Mountain View, CA) (see Supporting Materials and Methods for details).

Glossary

Abbreviations:

BM

bone marrow

BFU-E

burst-forming unit (erythroid)

CFU

colony-forming unit

FA

Fanconi anemia

MMC

mitomycin C

NAC

N-acetyl cysteine.