KIT blockade is sufficient for donor hematopoietic stem cell engraftment in Fanconi anemia mice

To the editor:

Fanconi anemia (FA) results from defects in genes involved in the DNA repair pathway and is characterized by progressive bone marrow failure (BMF) and a high incidence of cancer.¹  Currently, hematopoietic stem cell transplant (HSCT) is the only curative option for the BMF. Due to the underlying DNA repair defect, FA patients have inherent hypersensitivity to agents like cyclophosphamide, busulfan, and ionizing radiation.²  Recently, fludarabine, reduced-intensity cyclophosphamide, total body irradiation (TBI)–based conditioning followed by a CD34-selected graft has improved the 5-year survival to >80%.^3,4  Despite reduced graft rejection and graft-versus-host disease, conditioning-related toxicity and long-term risk of malignancy are still high.^3-6  Pretransplant conditioning devoid of alkylating agents/TBI would decrease the conditioning-related toxicity and additionally lower the risks of long-term malignancy in FA.

Unlike humans, most murine models of FA do not develop overt BMF, but have subtle defects in hematopoietic stem and progenitor cells (HSPCs),^7,8  and BMF becomes apparent upon experimentally induced hematopoietic stress.^8,9  Cytokinesis failure,¹⁰  inflammatory cytokine-driven apoptosis,⁹  and exhaustion of hematopoietic stem cells (HSCs) with replication-induced accumulation of DNA damage are some of the proposed mechanisms underlying the BMF in FA.¹¹ 

A prior study targeted at developing non–chemotherapy conditioning exploited the hypersensitivity of FA HSCs to the proinflammatory cytokine interferon-γ (IFN-γ); continuous exposure to IFN-γ facilitated donor HSC engraftment in FA mouse models.¹²  However, its clinical translation has been limited, likely due to findings of leukemic clonal evolution following exposure to inflammatory cytokines in FA.^13,14  Recently, blocking stem cell factor (SCF) binding to its cognate receptor KIT, via a monoclonal antibody (Ab) ACK2 (KIT-Ab), was shown to facilitate donor HSC engraftment in immune-deficient mice.¹⁵  However, no appreciable donor engraftment occurred in immunocompetent mice unless low-dose irradiation or CD47 blockade was used with ACK2.^16,17  Herein, we show that in FA, due to inherent HSC defect, targeted KIT-blockade alone results in significant normal donor engraftment following HSCT.

Eight to 12-week-old Fanca^−/−, Fancd2^−/−, and C57BL/6J mice were analyzed for frequency of HSPC before and 7 days following ACK2 injections, and the cohorts receiving ACK2 were subjected to definitive HSCT with 20 × 10⁶ or 5 × 10⁶ whole bone marrow (BM) cells from wild-type (WT; B6.SJL-Ptprc^a Pepc^b/BoyJ CD45.1) donors. Because FA mice were not fully backcrossed to the B6 strain and were not congenic, additional HSCTs were done following ACK2 and immunosuppression with a CD4-depleting Ab, GK1.5 (supplemental Figure 1, available on the Blood Web site).¹⁸  Supplemental “Materials” provide methodological details.

Although total BM cellularity was similar between WT and Fanca^−/− mice, the absolute numbers of HSPCs (Lineage⁻Sca⁺Kit⁺ [LSK]) and the more primitive LSK–signaling lymphocyte activation molecule (SLAM) cells (the “HSC compartment”) were significantly decreased in Fanca^−/− mice (Figure 1A-C). We hypothesized that attrition of HSPC would cause replicative stress and vice versa. Indeed, higher numbers of Fanca^−/− HSPC were in cycle compared with WT HSPC (Figure 1D), consistent with prior reports.¹⁹ 

We noted that Fanca^−/− HSPC had significantly reduced KIT surface expression in both LSK cells (Figure 1E-F) and the lineage-negative side population cells (supplemental Figure 1A-B). However, total KIT expression (surface and intracellular) was similar in WT and FA mice, suggesting more KIT internalization following SCF binding in Fanca^−/− HSPC (Figure 1F), and expression of SCF was upregulated in Fanca^−/− BM stroma (Figure 1G). Its levels trended higher in BM supernatant (supplemental Figure 1C) compared with WT counterparts, consistent with a report in FA patients.²⁰  Also, there was a trend to increased pERK signaling following SCF in FA LSK cells compared with WT controls (supplemental Figure 1D). Taken together, these finding suggested increased KIT signaling in Fanca^−/− HSPCs.

Indeed, Fanca^−/− LSK cells showed a proapoptotic signature at baseline by messenger RNA deep sequencing (mRNAseq), which was remarkably augmented with exposure to KIT-Ab ex vivo, in sharp contrast to WT LSK cells (Figure 1H). In addition, there was a marked downregulation of expression of cell-cycle and DNA repair response genes (supplemental Figure 2A-B). Functionally, this translated to pronounced apoptosis and cell death in Fanca^−/− LSK cells, compared with WT following ACK2 exposure (Figure 1I-K). Treatment with imatinib, a BCR-ABL tyrosine-kinase inhibitor that also blocks downstream KIT signaling, also increased apoptosis in Fanca^−/− LSK cells (supplemental Figure 1E), although the combination was not additive/synergistic, suggesting they both targeted the same KIT-signaling pathway (supplemental Figure 1F). In addition, 7 days following KIT-Ab injection, lineage-negative cells and colony forming units in the BM of Fanca^−/− and Fancd2^−/− mice were significantly decreased, when compared with similarly treated WT controls (Figure 2A-C). Hence, KIT blockade resulted in increased HSPC depletion both ex vivo and in vivo.

To determine whether KIT-Ab would suffice as a sole pretransplant conditioning regimen in Fanca^−/− and Fancd2^−/− mice, HSCTs were performed following KIT-Ab (supplemental Figure 3), with 22% to 24% donor/WT engraftment at 4 months in Fancd2^−/− and Fanca^−/− recipients, respectively (supplemental Figure 4A-B). Engraftment progressively increased in primary recipients. When secondary transplants were performed, there was a remarkable increase (∼83%) in WT hematopoiesis (supplemental Figure 4C-D), suggesting a strong selective advantage to WT HSC from the regenerative proliferation-stress.

We also discovered that the WT CD45.1 donor mice (B6.SJL-Ptprc^a Pepc^b/BoyJ) Fanca^−/− and Fancd2^−/− C57 CD45.2 recipients had minor mismatches in their genetic background. Congenic strains are usually a 99.8% genetic background match. A single-nucleotide polymorphism analysis showed that Fanca^−/− and Fancd2^−/− recipient mice were only 96% to 97% matched to donor B6.SJL-Ptprc^a Pepc^b/BoyJ mice (supplemental Table 1). Therefore, there could be minor allele mismatches that contribute to the engraftment barrier. Hence, we additionally depleted CD4 cells with the GK1.5-Ab along with ACK2.¹⁸  Also, depletion of CD4 T-regulatory cells has been proposed to enhance Ab-dependent cell-mediated cytotoxicity.¹⁷  HSCT using this approach showed higher primary multilineage donor engraftment that progressively increased to 63% and 93% in Fanca^−/− and Fancd2^−/− primary recipients, respectively (Figure 2D-G). ACK2 alone or in combination with GK1.5 Ab led to a progressively increasing WT graft, although the combination was superior (supplemental Figure 4E). Increased WT engraftment in peripheral blood was associated with selection seen at the level of HSC and progenitors in BM (Figure 2H-I). In addition, KIT-Ab was well tolerated with a patchy depigmentation of fur and transient mild weight loss noted (supplemental Figure 5).

Collectively, herein we show that FA HSPCs are susceptible to KIT-Ab–mediated apoptosis from increased KIT signaling, and this finding can be capitalized to develop an HSCT conditioning regimen. KIT-Ab conditioning can achieve predominantly donor hematopoiesis in both Fanca^−/− and in Fancd2^−/− mice. Our observation of progressively increasing normal hematopoiesis with time confirms the distinct selective advantage of normal HSPC at physiological steady state. The similar selective advantage is also noted in FA patients with revertant mosaicism in the lymphohematopoietic compartment; when reversion occurs in HSPCs, it prevents overt BMF.^21-24 

We used 20 × 10⁶ donor BM cells/mouse or 7.5 × 10⁸ total nucleated cells (TNCs)/kg, which is within the 5 to 10 × 10⁸ TNCs/kg targeted in most reduced-intensity HSCTs in humans.²⁵  Even 5 × 10⁶ total BM cells/mouse (2 × 10⁸TNCs/kg) resulted in progressively increased engraftment to 15% by 16 weeks with ACK2 (supplemental Figure 4F).

Despite replacement of the majority of FA hematopoiesis with normal in FA mice, the remainder of FA HSPCs can potentially evolve into myelodysplastic syndrome (MDS)/acute myeloid leukemia (AML). However, we observed no evidence of MDS/AML evolution in any of the transplanted FA mice despite a year-long follow-up.

In summary, a nongenotoxic conditioning regimen is a critical unmet need in FA patients. Despite current reduced-intensity potentially DNA-damaging conditioning,^3,6  there is significant associated morbidity and augmented cancer susceptibility in FA. KIT-targeted conditioning, shown herein, is clinically translatable and could markedly decrease the conditioning-associated toxicity and improve the short- and long-term HSCT outcomes in patients with FA.