Antimetabolites synergize with non-genotoxic antibody drug conjugate conditioning in hematopoietic stem cell lentiviral gene therapy

Abstract

Hematopoietic stem cell transplantation (HSCT) of genetically engineered cells is a promising treatment modality for monogenic diseases. However, hematopoietic stem cell (HSC)-directed lentiviral vector (LV) gene therapies remain limited by genotoxicities associated with standard non-targeted conditioning using irradiation or alkylating chemotherapy. We and others developed an antibody drug conjugate (ADC) with anti-CD117 and saporin (sap), which selectively depletes CD117-expressing HSC and progenitor cells (HSPCs). Since anti-CD117-sap does not generally provide engraftment comparable to conventional conditioning, we investigated the hypothesis anti-CD117-sap is not effective as a single conditioning agent due to insufficient HSPC depletion. We show anti-CD117-sap induces nearly complete residual HSPC entry into a non-G₀ proliferative state and postulated agents targeting cycling cells would enhance engraftment. When we administered the antimetabolite 5-fluorouracil with anti-CD117-sap, anti-thymocyte globulin, and an immunoglobulin-degrading enzyme, we achieved >90% peripheral blood chimerism of unmodified cultured donor cells. Additional immunosuppression combined with an increase in transplanted cells further permitted ∼70% engraftment of HSPCs transduced with an LV encoding a high-expression, immunogenic factor VIII (FVIII) transgene. Copy numbers of ∼1.0 and normal FVIII plasma activity levels (>50%) by 6 months post-HSCT were achieved. Collectively, we show ADC and antimetabolite administration can augment LV-engineered HSPC engraftment by targeting cycling cells.

Graphical abstract

Okalova and colleagues show deep depletion of bone marrow HSCs is not achieved with non-genotoxic antibody drug conjugates, like anti-CD117-saporin, but this agent does induce cycling of residual HSCs and progenitor cells. Repurposing antimetabolites, such as 5-fluorouracil, that target cycling cells induces deep depletion and augments engraftment of gene-modified cells.

Introduction

Hematopoietic stem cells (HSCs) are a rare, self-renewing population in the bone marrow (BM) that support homeostasis by differentiating into a lifelong supply of blood and immune cells. Classically, HSCs can be defined as either long-term (LT) or short-term (ST) depending, in general, on their CD34 expression and signaling lymphocytic activation molecule (SLAM) profile. The CD34− and SLAM+ (CD48− CD150+) LT-HSCs are a quiescent population with LT reconstitution capacity between 3 and 4 months.¹ The CD34+ and SLAM− (CD48− CD150−) ST-HSCs offer only transient peripheral blood reconstitution (<1 month).¹ Extrinsic cytokines and intrinsic transcription factors control LT-HSC differentiation into ST-HSCs, which then differentiate into multipotent progenitors (MPPs; CD48+ CD150−) and into mature myeloid or lymphoid lineages. Based on these self-renewal and multipotent properties, HSCs are an ideal gene therapy target using lentiviral vector (LV)-based strategies. Specifically, hematopoietic stem cell transplantation (HSCT) of LV-modified hematopoietic stem and progenitor cells (HSPCs) has shown great promise in the treatment of monogenic diseases. However, this treatment modality remains limited by two main genotoxic issues, (1) incidence of genotoxicities associated with the conventional conditioning required prior to transplantation of genetically modified cells and (2) vector insertion site mutagenesis. The use of self-inactivating LVs has dramatically reduced the risks associated with transgene integration in HSCs. However, there remains a major need for less conditioning-related genotoxic regimens.

Current pre-transplantation conditioning utilizes total body irradiation (TBI) or alkylating chemotherapeutics like busulfan, treosulfan, or melphalan in order to deplete cells in the BM niche, which creates “space” for, and stimulates, the engraftment of incoming HSCs.^2,3,4 Conventional conditioning is also often combined with immunosuppression, such as fludarabine, and prevents an immune response against the transplanted cells or against highly immunogenic transgene products in the context of gene therapy. Unfortunately, since pre-transplantation conditioning is non-targeted and DNA damaging, it is associated with harmful toxicities and genotoxicity-related potential infertility and malignancy. Therefore, the development of a clinically translatable non-genotoxic conditioning regimen is a high priority, translational objective that is needed for the general advancement of HSC-directed gene therapies for hematological disorders, and specifically for hemophilia A (HA).

HA is a monogenic congenital bleeding disorder affecting nearly 1/5,000 male births and is caused by a mutation in the F8 gene, which results in a deficiency in the blood coagulation factor VIII (FVIII) activity. For example, in severe cases (<1% FVIII activity) patients are at an elevated risk of spontaneous life-threatening hemorrhages into closed areas like the intracranial or intra-abdominal cavities. Prophylaxis FVIII replacement, FVIII mimetic therapy, or non-factor therapeutics that encompass mimetics and rebalancing agents are currently the current standard of care. However, these are complicated by the development of treatment-related neutralizing anti-FVIII IgG (termed inhibitors) in a high proportion of people with HA.⁵ These unmet clinical needs make HA a possible candidate for HSCT-directed gene therapy. Our laboratory has optimized a bioengineered high expression FVIII variant, termed ET3, which is utilized for recombinant LV delivery to HSPCs.^6,7,8,9,10 Expression of the ET3 transgene is driven by a CD68 promoter restricting expression to monocyte and macrophage-lineages upon HSPC differentiation. This CD68-ET3-LV has shown promise in a phase 1 clinical trial, where five subjects with severe HA were transplanted with transduced autologous HSCs, resulting in stable FVIII expression and activity correlating to vector copy number (VCN) in the peripheral blood.¹¹ Since this treatment modality has already shown proof of concept for HSPC-directed LV gene therapy for HA, the replacement of the conventional conditioning with an alternative, non-genotoxic strategy could significantly advance the application of this treatment strategy.

There are various non-genotoxic conditioning platforms currently being assessed to achieve targeted HSPC depletion in the fields of HSCT and HSC-directed gene therapy. Since the early 2000s, there have been several pioneering preclinical discoveries across different platforms including monoclonal antibodies; antibody drug conjugates (ADCs), chimeric antigen receptor T cells, and lipid nanoparticles.^{12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34} However, among these strategies, the ADCs show great potential to replace conventional conditioning since they deliver cytotoxic but non-genotoxic payloads to the cell types important for engraftment. For example, we and others in the field have developed an ADC by conjugating an anti-CD117 antibody to the type 1 ribosome inactivating toxin saporin (sap).^{24,25,26,27,29} After anti-CD117-sap is internalized by CD117-expressing HSPCs, the linker is cleaved in the lysosome releasing the cytotoxic sap payload into the cytosol to induce apoptosis. Anti-CD117-sap can achieve targeted depletion of the stem cell niche since this specific receptor tyrosine kinase plays a crucial role in HSPC survival, development, and signaling. However, in previous studies, we found anti-CD117-sap alone does not provide sufficient HSPC depletion to consistently achieve robust engraftment of CD68-ET3-LV-modified HSPCs that is comparable to engraftment using irradiation.²⁹ We, therefore, initiated studies to investigate the hypothesis that anti-CD117-sap is not effective as a single non-genotoxic conditioning agent in HSPC-directed LV gene therapy for HA due to a lack of deep HSPC depletion in the BM.

Studies show that even weak BM stress can reactivate dormant stem cells into a proliferative state.³⁵ Herein, we reaffirm this finding and show that anti-CD117-sap induces nearly complete residual HSPC exit from a quiescent G₀ state and entry into a non-G₀ proliferative state in mice. Therefore, to target these ADC-induced cycling cells, we expanded our conditioning regimen to include the proof of concept antimetabolite, 5-fluorouracil (5-FU), which exerts its cytotoxic effects on cells following phosphorylation into its active metabolites 5-fluorodeoxyuridine monophosphate (5-FdUMP), 5-fluorodeoxyuridine triphosphate (5-FdUTP), and 5-fluorouridine triphosphate (5-FUTP). Notably, 5-FU is well-recognized for its ability to induce cell-cycle arrest and apoptosis through 5-FdUMP and 5-FdUTP-mediated DNA replication arrest, activation of p53 signaling, inhibition of mitochondrial ATP synthesis, and production of reactive oxygen species.^36,37,38 However, it has recently been shown that 5-FUTP integration into RNA affects ribosome synthesis contributing significantly to the anti-proliferative activity of 5-FU. When we repurposed 5-FU into our non-genotoxic conditioning regimen, we showed that we are able to take advantage of 5-FU’s expansive mechanism of action and now target residual ADC-induced cycling stem cells.³⁹ The synergy seen between anti-CD117-sap and 5-FU not only augments engraftment of unmodified stem cells but also serves as a robust engraftment strategy for lentiviral gene-modified HSPCs, which was shown in HSPC-directed LV gene therapy for HA when combined with immune suppression. Consequently, these studies provide the basis for further investigation of the status of HSPC cycling post-ADC treatment and serve as proof of concept for repurposing small molecules into non-genotoxic conditioning regimens to augment engraftment of gene-modified cells, even in the context of a highly immunogenic transgene product.

Results

Residual progenitor populations remain after conditioning with optimized CD117 immunotoxin (anti-CD117-sap), limiting HSPC engraftment

We first tested the optimization of the ADC as a strategy to improve the efficacy of the CD117 immunotoxin (Figure 1A). Currently, the generation of our in-house anti-CD117-sap consists of two phases, which includes the initial linkage of streptavidin (strep) to sap, followed by subsequent conjugation of the strep-sap to a biotinylated CD117 mAb 2B8 clone (Figure S1A). As an optimization strategy, we characterized several variations using a strep linker from Prospec or Abcam manufacturers. We also tested the optimization of anti-CD117-sap (Figure S1B) by comparing commercial versus in-house biotinylation of our CD117 mAb and found these conjugates showed no differences in HSPC depletion (data not shown).

Figure 1.

In-house-developed anti-CD117-saporin (sap) immunotoxin fully depletes short-term (ST)/long-term (LT) HSCs but not the multipotent progenitors (MPPs), resulting in increased colony forming unit (CFU) assay counts compared to total body irradiation (TBI)

(A) Schematic of conjugation of in-house recombinant strep-sap to biotinylated CD117 monoclonal antibody (mAb) clone 2B8. Drug antibody ratio shown as 4 but is between 2 and 4 when we compare our in-house immunotoxin to a commercially conjugated version. (B) Coomassie-stained SDS-PAGE under reduced and non-reduced conditions shows polished sap, biotinylated CD117 mAb, streptavidin (strep)-sap utilizing strep from Prospec (PRO-283) manufacturer, and optimized CD117 immunotoxin (anti-CD117-saporin) with its associated conjugates. (C) Schematic depicting optimized CD117 immunotoxin pharmacokinetics assessment using a rat IgG ELISA kit to test amount of circulating immunotoxin in WT mice on days 3–7 following conditioning with 0.75 mg/kg. (D) Residual anti-CD117-sap (ng/mL) in diluted plasma was detected by anti-rat IgG ELISA with a rat IgG standard and shows clearance below the pooled naive plasma control by day 6 post-treatment. Control pooled naive plasma was 79.35 ng/mL. Data represent mean of sample (n = 2). (E) Schematic depicting optimized CD117 immunotoxin pharmacodynamics assessment by LSK-SLAM flow and HSPC-CFU assay 5 days after dosing 0.75 mg/kg in hemophilia A (HA) mice. (F) Representative flow cytometry pseudocolor plots depict MPP and ST/LT HSPC compartments in bone marrow (BM) leukocytes from conditioned mice. (G) LSK-SLAM quantification of HSPCs after conditioning with original versus optimized anti-CD117-sap compared to a naive negative control and TBI (800 RAD) positive control mouse. Data represent mean of sample (n = 2–4 per experimental group). (H) Colony numbers per 20,000 plated cells isolated from anti-CD117-sap conditioned mice (n = 1 per control group and 2–4 per experimental group). Colonies counted 10 days after being plated in triplicates in MethoCult GF M3434 medium. Triplicate colony count data represents mean ± sample SD per mouse.

Of the combinations tested, we show the most effective construct consisted of the Prospec strep-sap conjugated to the commercial biotinylated CD117 mAb, and this optimized candidate had comparable molecular weights on an SDS-PAGE gel (Figure 1B) to our previously published data.²⁹ After validating the conjugation by gel electrophoresis, the optimized anti-CD117-sap was administered to mice to assess the ADC’s plasma clearance, which used an anti-rat IgG ELISA that detected plasma anti-CD117 antibody (Figure 1C). These data are important because residual anti-CD117-sap has the potential to interfere with engraftment by targeting transplanted HSPCs. Six days after dosing 0.75 mg/kg of the optimized anti-CD117-sap the levels of residual circulating ADC were consistent with pooled naive control plasma levels at ∼80 ng/mL (Figure 1D). These levels confirm complete clearance of the optimized anti-CD117-sap and a pharmacokinetic profile similar to the original ADC previously administered on a 5-day conditioning timeline.²⁹ We did also observe 300 ng/mL of residual circulating anti-CD117-sap only 3 days after dosing, which is roughly a 30-fold decrease from levels at the time of administration. However, we show this low residual amount of our optimized ADC is not detrimental for engraftment and thus utilize timelines ranging between 3 and 5 days, as specified per each individual experiment, throughout the course of these studies.

Next, we investigated the pharmacodynamic profile of the optimized ADC by measuring HSPC depletion (Figure 1E), which was measured by lineage^-, Sca-1⁺, and CD117⁺ (LSK)-SLAM flow cytometry (gating strategy outlined in Figure S2A) and an HSPC-colony forming unit (CFU) assay. Anti-CD117-sap cleared the ST/LT-HSC compartments with the optimized ADC being comparable to irradiation (Figures 1F and 1G), and this trend is consistent regardless of whether BM was harvested through crushing or flushing femurs prior to flow analysis (Figure S2B). However, interestingly, the HSPC-CFU colony counts from individual mice were not decreased relative to our original, non-optimized ADC (Figure 1H). Although depleted, colony counts after ADC conditioning remain around half of the counts observed from naive untreated mice, whereas irradiation eradicates all counts (Figure S2C). From these studies, we show that even the optimized anti-CD117-sap only depletes ST/LT-HSCs, as determined by flow cytometry. However, the MPP population is not eradicated (Figure 1G), nor are the residual stem/progenitor cells that lead to colony formation as measured by CFU assays.

Anti-CD117-sap induces cycling in residual HSPCs

It has been well studied that hematopoietic stress can reactivate dormant stem cells and induce these cells into a cycling state.³⁵ Therefore, we next analyzed the cell cycle status of murine HSPCs from anti-CD117-sap conditioned mice. Using previously developed strategies to assess cell cycle status of low number stem cell populations, depicted in Figures S3A and S3B, we adapted our LSK-SLAM flow analysis to measure proliferation using Ki-67 and 4',6-diamidino-2-phenylindole (DAPI).⁴⁰ As illustrated in Figure 2A both wildtype (WT) CD45.1 and HA mice were conditioned with our optimized ADC at doses ranging from 0.5 to 1 mg/kg and sacrificed 24 h after treatment for BM collection and flow analysis.

Figure 2.

Anti-CD117-sap administration induces dose-dependent residual HSC and progenitor (HSPC) cell exit from a G₀ state and entry into a non-G₀, cycling state in wildtype (WT) and hemophilia A (HA) mice

(A) Schematic depicting assessment of cell cycle of murine HSPCs in WT CD45.1 versus HA mice conditioned with increasing anti-CD117-sap doses (0.5–1.0 mg/kg). BM isolated after 24 h and subject to LSK-SLAM flow cytometric analysis coupled together with Ki-67 and DAPI staining. (B) Representative flow cytometry plots look at Ki-67 versus DAPI in the gated MPP and ST/LT HSPC compartments where there is an increase in signal intensity of cycling Ki-67-positive cells (purple histogram) compared to quiescent Ki-67-negative cells (blue histogram) after anti-CD117-sap treatment. (C) Cycling flow quantification of residual gated MPP and ST/LT HSC compartments reveals anti-CD117-sap induces exit from quiescent G₀ and entry into a proliferative non-G₀ state (S/G₂/M) as determined by Ki-67 mean fluorescent intensity (MFI). Data represent mean of sample (n = 2 per control and experimental group). (D) Plots depict the mean percent of residual G₀ and S/G₂/M cycling cells gated from the MPP and ST/LT HSPC compartments revealing anti-CD117-sap induction of stem cell proliferation is reproducible with another anti-CD117-sap batch (n = 2 per control and experimental group).

Flow analyses of cells from a representative mouse from each anti-CD117-sap dose group are shown relative to a naive (negative control) and a 5-FU (positive control) treated mouse (Figure 2B). Indeed, 5-FU was used as a positive control as it has been shown to induce cell cycling in mouse BM and activate dormant HSPCs.^35,41 The WT mice were tested in order to validate whether any trends we observed were unique to HA mice, and we found the optimized anti-CD117-sap displays a dose-response of extensive ST/LT-HSC depletion as early as 24 h in both strains (Figure S3C). Interestingly, the Figure 2B flow analysis of murine HSPCs revealed anti-CD117-sap induces residual HSPC exit from a quiescent G₀ state and entry into a non-G₀ proliferative state, and the corresponding HSPC proportions are provided in Table S1. This cycling induction was also further evidenced by a decrease in absolute cell numbers of LSK-SLAM-gated HSPCs that did not stain positive for either proliferation marker as well as an increase in Ki-67 mean fluorescent intensity (MFI) compared to the naive negative controls (Figure 2C). Finally, this decrease in HSPCs in G₀ compared to those cycling in the S/G₂/M phase was also reproducible when the flow analysis was repeated with a different anti-CD117-sap batch (Figure 2D).

Antimetabolites that target cycling cells can augment anti-CD117-sap depletion of HSPCs

Based on these results, we incorporated the timed administration of antimetabolites after ADC treatment as a strategy to target the cycling HSPCs with a goal of achieving deeper HSPC depletion. We specifically tested the small molecule drugs, fludarabine (Flu) and 5-FU, which are both toxic to cycling cells.^39,42 In addition, Flu and 5-FU have been widely studied in the context of murine HSC biology or applied in murine HSCT models as a component of pre-transplantation conditioning regimens, leading us to the rational decision to test these in combination with the anti-CD117-sap.^6,41,43,44

Flu was incorporated into the anti-CD117-sap conditioning regimen at either a single dose of 200 mg/kg or a total dose of 90 mg/kg given over three consecutive days after the ADC administration in HA mice (Figure 3A). Inclusion of this antimetabolite into the ADC regimen was well tolerated and achieved deep HSPC depletion as determined by LSK-SLAM flow cytometry (Figures 3B and 3C). However, HSPC-CFU assays on BM from anti-CD117-sap and Flu-treated mice showed this conditioning does not decrease colony counts (Figure 3D).

Figure 3.

Rational screening of the antimetabolites fludarabine (Flu) and 5-fluorouracil (5-FU) with anti-CD117-sap conditioning reveals only 5-FU incorporation into the regimen augments LSK-SLAM depletion and decreases CFU counts

(A) Schematic depicting LSK-SLAM and HSPC-CFU screening of the antimetabolite Flu at total doses of either 90 or 200 mg/kg in a conditioning regimen with anti-CD117-sap (0.75 mg/kg) in HA mice. (B) Representative flow cytometry pseudocolor plots depict MPP and ST/LT HSPC compartments in BM from anti-CD117-sap/Flu conditioned mice. (C) LSK-SLAM quantification of HSPCs after conditioning with anti-CD117-sap and Flu at total doses either 90 or 200 mg/kg compared to a naive negative control and total body irradiated (TBI; 800 RAD) positive control mouse. Data represent mean of sample (n = 3 per experimental group). (D) Colony numbers per 20,000 plated cells isolated from anti-CD117-sap and Flu conditioned mice (n = 1 per control group and 3 per experimental group). Colonies counted 10 days after being plated in triplicates in MethoCult GF M3434 medium. Triplicate colony count data represents mean ± sample SD per mouse. (E) Schematic depicting LSK-SLAM and HSPC-CFU screening of the antimetabolite 5-FU at a total dose of 25 mg/kg in a conditioning regimen with anti-CD117-sap (0.75 mg/kg). (F) Representative flow cytometry pseudocolor plots depict MPP and ST/LT HSPC compartments in BM from anti-CD117-sap/5-FU conditioned mice. (G) LSK-SLAM quantification of HSPCs after conditioning with anti-CD117-sap and 5-FU in presence and absence of immunosuppression compared to a naive negative control and total body irradiated (TBI; 800 RAD) positive control mouse. Data represent mean of sample (n = 2 per experimental group). (H) Colony numbers per 20,000 plated cells isolated from anti-CD117-sap and 5-FU conditioned mice (n = 1 per control group and 2 per experimental group). Colonies counted 10 days after being plated in triplicates in MethoCult GF M3434 medium. Triplicate colony count data represents mean ± sample SD per mouse and show that only anti-CD117-sap, 5-FU and ATG conditioning depletes colony counts comparable to irradiated control.

With respect to our second tested antimetabolite (5-FU), a dose escalation study showed a maximum tolerated dose (MTD) of 50 mg/kg 5-FU when incorporated into our non-genotoxic conditioning regimen (Figure S4A). Although anti-CD117-sap can be tolerated at higher doses when it is administered alone as we showed in our LSK-SLAM cycling experiments, it has an MTD of 0.75 mg/kg when combined with 5-FU. Unlike Flu, 5-FU is not highly immunosuppressive, so we also included an anti-CD117-sap/5-FU group in combination with 30 mg/kg antithymocyte globulin (ATG; Figure 3E), as our goal is to use this regimen in an LV-HA strategy that is highly immunogenic.^{45,46,47,48,49,50,51,52,53} Not only is the MPP compartment now fully cleared to levels achieved with irradiation across all groups, but Figure 3G highlights the absolute cell counts of the entire LSK compartment are now eradicated compared to the naive HA control. However, we only observed a decrease in the colony counts in the group combining anti-CD117-sap and 5-FU with ATG immunosuppression (Figure 3H). It is further shown that ATG alone or in combination with just 5-FU does not clear HSPCs or decrease colony counts (Figures S4B–S4D). Therefore, only anti-CD117-sap/ATG/5-FU achieves robust HSPC depletion comparable to irradiation. This result was confirmed (Figure S4E), but the exact mechanism of synergy remains unknown, but it may be attributed to ATG’s promiscuous binding profile.

Anti-CD117-sap and 5-FU conditioning permits robust engraftment of unmodified cells

Based on this proof of concept results, we hypothesized that this ADC/antimetabolite conditioning prior to HSPC transplantation would, in turn, permit engraftment levels similar to irradiation. To test this, we used the anti-CD117-sap/ATG/5-FU conditioning regimen in an HSCT protocol with 2 × 10⁶ sex-matched, unmodified, Sca-1⁺ HSPCs, as depicted in Figure 4A. However, we have observed inconsistent engraftment outcomes with ATG in the setting our murine HSCT protocols due to its polyvalent antibody composition and promiscuous binding profile to functional antigens present on HSPCs, such as C-X-C chemokine receptor 4 (CXCR4).^54,55,56 Russell et al. confirmed by flow cytometric analysis substantial ATG binding to primitive and mature HSPC populations in BM, peripheral blood, and spleen tissues further highlighting the potential for residual circulating ATG at the time of transplant to bind incoming donor Sca-1⁺ HSPCs.²⁹ We have also repeated our murine transplants with anti-CD117-sap with ATG only and have shown inconsistent engraftment (data not shown) since ATG remains in circulation for ∼30 days. Since murine ATG is not identical in performance to human ATG agents, we also included the immunoglobulin G-degrading (IgG) enzyme of Streptococcus pyogenes (IdeS) in the regimen to cleave residual ATG at the time of transplant. In Figure S5A, we show that IdeS is a cysteine endopeptidase with a high degree of substrate specificity for all subclasses of human, and some classes of monkey, rabbit, and sheep IgG.^57,58,59 Since rabbit, but not rat-IgG, is a substrate of IdeS, in the context of our regimen, it only acts on ATG because the anti-CD117-sap immunotoxin features an anti-rat CD117 biotinylated monoclonal antibody. We validated IdeS functional depletion of ATG (Figures S5B and S5C), and show it effectively cleaves rabbit anti-mouse ATG in vivo after 24 h at a single dose as low as 20 U/g.

Figure 4.

Anti-CD117-sap/ATG/5-FU conditioning permits rapid and robust engraftment with unmodified cultured Sca-1⁺ HSPCs, but at the expense of survival in HA-E16 mice

(A) Schematic depicting HSCT in WT and HA mice with 2 × 10⁶ sex-matched unmodified Sca-1⁺ HSPCs following conditioning with anti-CD117-sap (0.75 mg/kg), ATG (30 mg/kg), 5-FU (25 mg/kg), and IdeS (20 U/g). (B) Kaplan-Meier survival analysis shows full survival in WT mice to the first engraftment check after conditioning and transplant with unmodified HSPCs (n = 6). (C) Chimerism flow quantification of total peripheral blood donor engraftment in transplanted WT mice. Data represent mean ± sample SD (n = 6). (D) Kaplan-Meier survival analysis shows mortality variances in HA mice to the first engraftment check after conditioning and transplant with unmodified HSPCs (n = 6). (E) Chimerism flow quantification of total peripheral blood donor engraftment in transplanted HA mice. Data represent mean ± sample SD (n = 6). (F) Plots depict effects to white blood cell (WBC), lymphocyte, monocyte, and neutrophil populations that were observed post-HSCT in WT and HA mice. The two dotted lines represent the upper and lower limit of the reference range for each population. Data represent mean ± sample SD (n = 6 for WT strain and 3 surviving mice for HA strain).

Using this conditioning regimen, male and female WT mice showed 100% survival at the first engraftment check of 2 weeks post-HSCT (Figure 4B), and these mice achieved ∼75% peripheral blood chimerism by 6 months (Figure 4C) as measured by the flow cytometry panels outlined in Figure S6. The rate of engraftment in the WT mice is slow and chimerism levels >50% were only seen after 4 months. Interestingly, when the same regimen was applied to HA mice, there was 50% mortality (Figure 4D). However, the surviving HA mice displayed rapid engraftment with levels as high as 60% by the first engraftment check and stabilization at ∼80% by 2 months (Figure 4E), similar to that observed in the irradiation controls (Figure S5D). Mice conditioned without IdeS have inconsistent engraftment, where 50% do not engraft, and mice treated with 5-FU alone fail to engraft (Figure S5D). The necessity of ATG in this regimen is further demonstrated in Figure S7, where only a single HA mouse conditioned with anti-CD117-sap and just 5-FU achieved >50% engraftment levels, while all other mice failed to engraft. Collectively, these transplant results align with our HSPC-CFU assays showing the synergy of anti-CD117-sap/ATG/5-FU.

Lastly, as predicted, the anti-CD117-sap and Flu conditioning regimen was unsuccessful when tested with both unmodified and gene modified cells. All the anti-CD117-sap/Flu-conditioned mice combined from the two transplant groups failed to engraft and their complete blood counts (CBCs) remained on the lower limit of the reference range for the duration of the experiment (Figures S8A–S8C). As an additional control, we also conducted an LSK-SLAM and HSPC-CFU assay screening to test if anti-CD117-sap/ATG/Flu display similar synergy as that observed with 5-FU, but we confirmed they do not as the conditioning did not fully deplete CFU counts (Figures S8D and S8E). Thus, we focused the remainder of our experiments on exploring the anti-CD117-sap/ATG/5-FU transplant-related outcomes in the HA strain in preparation for translating the regimen into a gene therapy protocol using engineered HSPCs.

White blood cell counts impact ADC partitioning and engraftment outcomes

Based on CBCs (Figure 4F), the anti-CD117-sap/ATG/5-FU-conditioned HA mice did develop severe neutropenia after transplant unlike the WT mice. Figure S9A further highlights these conditioned and transplanted mice are not anemic and have platelet levels that are within normal range. Furthermore, decreasing the 5-FU dose is able to reverse the neutropenia that we observed in the HA mice (Figure S9B). Interestingly, we noted that white blood cell (WBC) counts in the HA strain started low but approached those of WT mice after 6 months. Because the transplanted Sca-1⁺ cells were harvested from WT mice, this suggested a strain difference in the WBCs of WT versus HA mice.

To further investigate potential strain differences, BM and blood was collected from WT and HA mice and analyzed for HSPC-CFU and CBCs (Figure 5A). These data show our HA-E16 mice have higher CFU counts relative to the WT B6 strain (Figure 5B). Our E16 strain, which is a standard HA mouse model, has about 70% B6 background.⁶⁰ In addition to the E16 model, this strain has been extensively backcrossed with B6 mice, and these HA mice, termed MIRB, have >97% B6 background.⁶¹ These HA-MIRB mice retain a similar bleeding phenotype but show a less robust immune response to FVIII than observed in HA mice with the mixed background.⁵¹ Unlike the E16 strain, the HA-MIRB mice display both lower CFU colony counts and higher WBC counts that are comparable to WT mice strain (Figure 5C). We then repeated the anti-CD117-sap/ATG/5-FU conditioning regimen in the HA-MIRB mice and found 100% survival with this strain (Figures 5D and 5E). Based on these results, we also tested if (1) the addition of FVIII prophylaxis or (2) administration of splenic leukocytes to boost WBC levels could protect HA-E16 mice from the anti-CD117-sap/ATG/5-FU regimen. Indeed, both were able to reverse the deaths observed in the original HA-E16 strain and achieve 100% survival (Figure 5E). In addition, apart from one mouse in the leukocyte supplementation group, these mice engrafted at robust levels of >75% chimerism by 6 months post-HSCT (Figure 5F). The red blood cell (RBC), hemoglobin, and platelet levels were also comparable to what we observed previously in WT mice and the HA-E16 strain (Figure S9C). Across all three groups, the HA mice had WBCs counts that were in the normal range by 6 months post-HSCT (Figure 5G).

Figure 5.

HA-MIRB mice with white blood cell (WBC) counts comparable to WT mice tolerate 5-FU and anti-CD117-sap conditioning, but factor VIII (FVIII) prophylaxis or splenic leukocyte coverage are needed to overcome phenotypic consequences of transplant related deaths in HA-E16 mice with lower WBCs than WT mice

(A) Schematic illustrating blood and BM collection from naive WT, HA-MIRB, and HA-E16 mice before subsequent HSPC-CFU assay and CBC analysis, respectively. (B) Colony numbers per 20,000-plated cells isolated from naive WT, HA-MIRB and HA-E16 mice. Colonies counted 10 days after being plated in triplicates in MethoCult GF M3434 medium. Triplicate colony count data represents mean ± sample SD per mouse (n = 1–3 per strain). (C) Representative WBC counts from naive WT (GFP), WT (CD45.1), HA-MIRB, and HA-E16 mice. The dotted line represents the lower limit of reference range for each population. Data represent mean ± sample SD (n = 5–18 per strain). (D) Schematic depicting HSCT in different HA mouse strains with 2 × 10⁶ sex-matched unmodified Sca-1⁺ HSPCs following conditioning with anti-CD117-sap (0.75 mg/kg), ATG (30 mg/kg), 5-FU (25 mg/kg), and IdeS (20 U/g). HA-MIRB received no additional coverage, whereas the HA-E16 mice received either FVIII prophylaxis 24 h before conditioning with 2 μg of FVIII or injection with 20 × 10⁶ red blood cell lysed splenic leukocytes. (E) Kaplan-Meier survival analysis shows full survival in HA-MIRB mice with just the regimen alone and HA-E16 mice with FVIII prophylaxis or splenic leukocytes to the first engraftment check after conditioning and transplant with unmodified HSPCs (n = 3–4 per experimental group). (F) Chimerism flow quantification of total peripheral blood donor engraftment in transplanted HA-MIRB and HA-E16 mice. Data represent mean ± sample SD (n = 3–4 per experimental group). (G) Plots depict effects on WBC, lymphocyte, monocyte, and neutrophil populations that were observed post-HSCT in HA-MIRB and HA-E16 mice. The two dotted lines represent the upper and lower limit of the reference range for each population. Data represent mean ± sample SD (n = 3–4 per experimental group).

Anti-CD117-sap and 5-FU conditioning is not sufficiently immunosuppressive for engraftment of modified cells expressing FVIII

Because WBCs appear to be a marker for sensitivity and tolerance to the conditioning regimen, we further tested the various strains in an HA-LV gene therapy protocol. Sca-1⁺ cells were transduced at a multiplicity of infection (MOI) of 5 twice with an LV encoding our high-expression FVIII transgene termed ET3, with a myeloid-directed CD68 promoter (Figure 6A), resulting in a VCN from CFUs of ∼3 (Figure S10). After transplant, all HA groups had lower total engraftment with gene-modified cells compared to engraftment with non-modified cells (Figure 6B). Not only was engraftment lower but mice also displayed a high degree of variability in donor engraftment levels across myeloid and non-myeloid cell types, as determined by an expanded chimerism flow analysis panel (Figure S11). Transgene expression of FVIII is immunogenic, and Figure 6C shows the decreased and variable engraftment is associated with individual mice developing varying levels of anti-ET3 IgG across all three groups. With respect to the HA-E16 mice that received FVIII prophylaxis, all 4 mice developed anti-ET3 IgG, ranging between titers of 13 and 427. A similar trend of anti-ET3 IgG formation was observed in HA-E16 mice that received splenic leukocytes. In contrast, 2 of 4 mice in the HA-MIRB group developed anti-ET3 IgG. The mice that did not form anti-ET3 IgG had peripheral blood VCNs between 2 and 3 and FVIII plasma levels increased from 5% normal levels at 2 weeks to over 90% by 6 months (Figures 6D and 6E). Interestingly, these mice also displayed increases in FVIII activity that occurred around 4 months post-transplant, while also achieving total engraftment levels between 50% and 70% by 6 months.

Figure 6.

Conditioning with 5-FU and anti-CD117-sap is sufficiently HSPC depleting, but the regimen is not immunosuppressive enough in the context of a HA LV gene therapy protocol, as evidenced by the formation of anti-ET3 IgG

(A) Schematic depicting HSCT with 2 × 10⁶ codon optimized CD68-ET3-LV transduced Sca-1⁺ HSPCs at an MOI of 5 × 2 in HA-MIRB or HA-E16 mice following conditioning with anti-CD117-sap (0.75 mg/kg), ATG (30 mg/kg), 5-FU (25 mg/kg), and IdeS (20 U/g). HA-MIRB receives no additional coverage, while HA-E16 mice receive either WBC or FVIII coverage 24 h prior to anti-CD117-sap dosing. (B) Chimerism flow quantification of CD45.1 (%) engraftment per cell type in HA-MIRB and HA-E16 mice transplanted with 2 × 10⁶ codon optimized CD68-ET3-LV cells (n = 3–4 per experimental group). (C) Anti-ET3 IgG ELISA analysis on plasma from HA-MIRB and HA-E16 mice transplanted with 2 × 10⁶ codon optimized CD68-ET3-LV cells (n = 3–4 per experimental group). (D) Vector copy number was determined by quantitative PCR of proviral DNA from blood collected at 12 weeks following transplant in HA-MIRB and HA-E16 mice with 2 × 10⁶ codon optimized CD68-ET3-LV cells. Data represent mean of sample (n = 3–4 per experimental group). (E) ET3 activity levels as determined by chromogenic assay on plasma collected from HA-MIRB and HA-E16 mice transplanted with 2 × 10⁶ codon optimized CD68-ET3-LV cells (n = 3–4 per experimental group).

From these studies, it was concluded that the two primary issues still to be elucidated and overcome were (1) anti-ET3 IgG formation against the transgene product and (2) the lower engraftment levels of gene-modified cells compared to what we observed with unmodified cells. One concern was the use of a high expression FVIII transgene, which was codon-optimized for high level expression. To first address the anti-ET3 IgG formation, we chose to instead test our CD68-ET3-LV featuring a non-codon optimized FVIII transgene and transplanted HA-MIRB mice (Figure 7A). While these mice had less variable engraftment, the chimerism levels were still substantially lower than in the mice transplanted with untransduced cells (Figure 7B). Even though the non-codon-optimized construct expressed at lower levels, these mice still developed anti-ET3 IgG titers during the first 2 months that did decrease by 6 months post-HSCT (Figure 7C). However, they only had a small increase in ET3 activity that was still <1% despite a mean VCN of ∼2 (Figures 7D and 7E). Therefore, it was apparent from these results that immunogenicity stemming from ET3 expression was limiting LT engraftment without additional immunosuppression.

Figure 7.

Transduction itself seems to decrease ability of cultured HSPCs to engraft following 5-FU and anti-CD117-sap conditioning, and formation of anti-ET3 IgG persists across different CD68-ET3-LVs

(A) Schematic depicting HSCT with 2 × 10⁶ naive untransduced or non-codon optimized CD68-ET3-LV transduced cultured Sca-1⁺ HSPCs at an MOI of 25 × 2 in HA-MIRB only following conditioning with anti-CD117-sap (0.75 mg/kg), ATG (30 mg/kg), 5-FU (25 mg/kg), and IdeS (20 U/g). (B) Chimerism flow quantification of CD45.1 (%) engraftment per cell type in HA-MIRB mice with 2 × 10⁶ naive untransduced or non-codon optimized CD68-ET3-LV cells. Data represent mean ± sample SD (n = 4 for unmodified and 7 for gene-modified group). (C) Anti-ET3 IgG ELISA analysis on plasma from HA-MIRB mice transplanted with 2 × 10⁶ naive untransduced or non-codon optimized CD68-ET3-LV cells (n = 4 for unmodified and 7 for gene-modified group). (D) Vector copy number was determined by quantitative PCR of proviral DNA from blood collected at 12 weeks following transplant in HA-MIRB mice and 1 × 10⁶ naive untransduced or non-codon optimized CD68-ET3-LV cells. Data represent mean of sample (n = 4 for unmodified and 7 for gene-modified group). (E) ET3 activity levels as determined by chromogenic assay on plasma collected from HA-MIRB transplanted with 2 × 10⁶ naive untransduced or non-codon optimized CD68-ET3-LV cells. Data represent mean of sample (n = 4 for unmodified and 7 for gene-modified group). (F) Schematic depicting HSCT with 2 × 10⁶ codon optimized CD68-ET3-LV transduced Sca-1⁺ HSPCs at an MOI of 5 × 2 in HA-MIRB mice only following anti-CD117-sap, ATG, and IdeS conditioning but with an increased dose of 5-FU at either 37.5 of 50 mg/kg. (G) Chimerism flow quantification of CD45.1 (%) engraftment per cell type in transplanted HA-MIRB mice under increased 5-FU doses comparing gene-modified cells to mock transduced cells. Data represent mean ± sample SD (n = 3 per experimental group). (H) Anti-ET3 IgG ELISA analysis on plasma from HA-MIRB mice transplanted with 2 × 10⁶ codon optimized CD68-ET3-LV or mock transduced cells (n = 3 per experimental group). (I) Vector copy number was determined by quantitative PCR of proviral DNA from blood collected at 12 weeks following transplant in HA-MIRB mice with 2 × 10⁶ codon optimized CD68-ET3-LV or mock transduced cells. Data represent mean of sample (n = 3 per experimental group). (J) ET3 activity levels as determined by chromogenic assay on plasma collected from HA-MIRB mice transplanted with 2 × 10⁶ codon optimized CD68-ET3-LV or mock transduced cells (n = 3 per experimental group).

To test the concern of low engraftment, we sought to increase the 5-FU dose in the HA-MIRB strain since we already established mice with a higher B6 background are more tolerant to the combined regimen. HA-MIRB mice were conditioned with the full regimen, but at an increased 5-FU dose of either 37.5 or 50 mg/kg (Figure 7F). Both mock and codon-optimized CD68-ET3-LV-transduced cells were transplanted, which showed the myeloid and non-myeloid engraftment was comparable across all cell types (Figure 7G). However, even with mock transduction, the engraftment levels do not reach the robust levels that were achieved with unmodified cells using the anti-CD117-sap/ATG/5-FU conditioning, suggesting the transduction protocol interferes with engraftment. Furthermore, over half of the mice between the two gene-modified groups developed anti-ET3 IgG (Figure 7H). Despite having a mean VCN of ∼1, only 3 mice showed FVIII activity, which was variable and ranged between 12% and 60% by 32 weeks after HSCT (Figures 7I and 7J).

Collectively, these results confirm that anti-CD117-sap/ATG/5-FU conditioning is not sufficiently immunosuppressive to eradicate anti-ET3 IgG against the ET3 transgene. We also show that engraftment is inherently lower with gene-modified cells compared to unmodified cells even when the dose of antimetabolite is increased, and this is observed with mock-transduced cells, suggesting that the process of genetic modification using our protocol/LV impedes engraftment in non-genotoxic conditioning settings. Although the exact mechanism of interference following the process of modification is unknown, we speculated it could be a downstream implication of culturing the Sca-1⁺ HSPCs. For example, we utilize polybrene to increase the efficiency of gene transfer, which could be at the expense of toxicity in sensitive cell types like HSPCs.

Additional immunosuppression and increased cell number are needed with 5-FU and anti-CD117-sap conditioning to achieve robust engraftment of gene-modified cells

We previously tested a non-genotoxic conditioning regimen in HA-E16 mice utilizing a monoclonal antibody cocktail directed at cell surface antigens CD4, CD8, and CD40L.²⁹ This protocol achieved suppression of anti-FVIII immunity but did not result in high chimerism levels, and the myeloid and non-myeloid cell engraftment levels were variable. We, therefore, next tested the addition of antibody-directed immunosuppression into our conditioning regimen using HA-MIRB mice (Figure 8A). Engraftment levels were still low with an average peripheral blood engraftment of ∼25% at the final engraftment check (Figure 8B). However, the anti-CD117-sap/ATG/5-FU conditioning regimen with the immunosuppressive cocktail did eliminate anti-ET3 IgG formation (Figure 8C). These results suggest the addition of immunosuppression is not sufficient to overcome the low engraftment barrier of gene-modified cells, as further evidenced by a mean ET3 activity of only ∼15% at 24 weeks after HSCT, but is sufficient to eliminate anti-transgene product immunity (Figure 8D).

Figure 8.

Increasing immunosuppression and cell number are both needed to overcome anti-ET3 IgG and achieve high engraftment of gene-modified cells with corrected FVIII activity in a HA LV gene therapy protocol following 5-FU and anti-CD117-sap conditioning

(A) Schematic depicting HSCT with 2 × 10⁶ codon optimized CD68-ET3-LV transduced cultured Sca-1⁺ HSPCs at an MOI of 5 × 2 in different HA-MIRB mice following conditioning with anti-CD117-sap (0.75 mg/kg), ATG (30 mg/kg), 5-FU (50 mg/kg), IdeS (20 U/g), and antibody cocktail (CD4 + CD8 + CD40L mAbs). Anti-CD4 (200 μg), anti-CD8 (200 μg), and anti-CD40L (500 μg) were all dosed on the day of transplant. (B) Chimerism flow quantification of CD45.1 (%) engraftment per cell type in HA-MIRB mice under additional immunosuppression transplanted with 2 × 10⁶ codon optimized CD68-ET3-LV cells. Data represent mean ± sample SD (n = 5). (C) Anti-ET3 IgG ELISA analysis on plasma from HA-MIRB mice under additional immunosuppression transplanted with 2 × 10⁶ codon optimized CD68-ET3-LV cells (n = 5). (D) ET3 activity levels as determined by chromogenic assay on plasma collected from HA-MIRB under additional immunosuppression transplanted with 2 × 10⁶ codon optimized CD68-ET3-LV cells. Data represent mean ± sample SD (n = 5). (E) Schematic depicting HSCT with 6 × 10⁶ codon optimized CD68-ET3-LV transduced Sca-1⁺ HSPCs at an MOI of 5 × 2 in HA-MIRB mice only following anti-CD117-sap, ATG, 5-FU, IdeS, and antibody cocktail conditioning. (F) Chimerism flow quantification of CD45.1 (%) engraftment per cell type in HA-MIRB mice under additional immunosuppression transplanted with 6 × 10⁶ codon optimized CD68-ET3-LV cells. Data represent mean ± sample SD (n = 4 for HSCT 1 group and 6 for HSCT 2 group). (G) Anti-ET3 IgG ELISA analysis on plasma from HA-MIRB mice under additional immunosuppression transplanted with 6 × 10⁶ codon optimized CD68-ET3-LV cells (n = 4 for HSCT 1 group and 6 for HSCT 2 group). (H) Vector copy number was determined by quantitative PCR of proviral DNA from blood collected at 12 weeks following transplant in HA-MIRB mice under additional immunosuppression and 6 × 10⁶ codon optimized CD68-ET3-LV cells. Data represent mean ± sample SD (n = 10 for combined HSCT 1 and 2 groups). (I) ET3 activity levels as determined by chromogenic assay on plasma collected from HA-MIRB under additional immunosuppression transplanted with 6 × 10⁶ codon optimized CD68-ET3-LV cells. Data represent mean ± sample SD (n = 4 for HSCT 1 group and 6 for HSCT 2 group).

It is well recognized that increasing cell number can augment engraftment outcomes in non-genotoxic conditioning settings.^14,25,34 Because of the immunogenic barrier of FVIII-expressing transgenes, we next coupled the addition of our antibody cocktail into a transplant regimen with an increase from 2 × 10⁶ to 6 × 10⁶ gene-modified HSPCs (Figure 8E). The experiment was conducted across two separate cohorts in which, we transplanted 4 mice in the first HSCT and then repeated the experiment in a second HSCT with 6 additional mice. This increased cell number strategy successfully restored engraftment levels in both cohorts of mice toward the robust levels observed with unmodified cells. Specifically, there is an average chimerism level of ∼70% reached by 6 months post-HSCT in the first cohort of transplanted mice (Figure 8F). On the other hand, mice from the second HSCT showed lower total and myeloid engraftment at ∼45% and 50%, respectively (Figure 8F). Nonetheless, none of the mice across both transplant experiments developed anti-ET3 IgG and they had an average VCN ∼1.0 at 12 weeks post-HSCT (Figures 8G and 8H). Furthermore, all 10 mice, irrespective of engraftment level, had corrected FVIII activity levels in the normal range (>50%) by 4 months after HSCT (Figure 8I), with 100% survival (10/10). In both groups, we observed a sharp increase in activity 3 months after HSCT, in which activity increased from under 50% to over 70%. Interestingly, half of the mice in the second cohort increased to over a 1 U/mL ET3 activity despite showing slightly lower engraftment levels across all cell types compared to the first transplanted cohort. This trend of increasing activity is unique to this non-genotoxic conditioning regimen and has not been observed in the irradiation controls from previous studies.²⁹ It should also be noted that all mice displayed blood counts in the normal range throughout the course of the transplant apart from higher WBCs being observed in the mice from the second cohort (Figure S12A). In addition, this conditioning regimen did not impact the weight of these mice or negatively impact reproductive capacity, as evidenced by normal litter sizes from breeding pairs established following HSCT (Figures S12B and S12C).

Discussion

We demonstrated how conditioning with non-genotoxic conditioning ADCs, such as anti-CD117-sap, causes cycling of residual HSPCs in the BM niche. The cycling status of these residual HSPCs has the potential to impede engraftment of transplanted donor cells, a phenomenon that can be exacerbated in cases where genetically modified cells are being transplanted. We further established the potential of repurposing antimetabolites into these non-genotoxic conditioning regimens when combined with immune suppression to further target ADC-induced cycling HSPCs and augment depletion in the BM prior to HSCT. In this study, we also report HA strain variances that impact transplant outcomes with our specific conditioning regimen, indicating flow-based assays showing HSPC depletion are not always reliable predictors of engraftment outcomes in gene therapy settings. We found that even with optimized ADC and 5-FU dosing, this conditioning regimen must be combined with additional immunosuppressive reagents to be successful in an HA LV gene therapy protocol, where the FVIII transgene product is highly immunogenic.

We first optimized the CD117 ADC by testing various strep linkers, biotinylating strategies, and dosing schedules in an effort to improve efficacy. We found that conjugating a different strep source to the same biotinylated anti-CD117 we have previously used in linking our ADC was able to enhance total HSPC depletion. This highlights how slight differences in reagent selection for ADC production can have considerable implications on the conditioning pharmacodynamics. We also found that LSK-SLAM and HSPC-CFU assays were critical parameters to measure during the optimization of anti-CD117-sap ADC. Together, the combined screening tests were able to consistently predict transplant outcomes. When a conditioning regimen resulted in complete depletion in both assays, transplanted mice displayed robust engraftment with unmodified cells at levels comparable to irradiated positive controls. However, these assays did not translate to the transplantation of ET3-modified cells, and this could be due to the transduction process affecting the functionality of the HSPCs.^61,62,63,64 We ultimately had to provide additional immunosuppression and increase the number of transplanted cells to overcome the low engraftment barrier following transduction. Specifically, several studies show that the strategy of increasing cell number is sufficient to overcome low engraftment barriers after non-genotoxic conditioning, resulting in increased LT engraftment of stem cells.^14,25,34 When applied in our HA LV gene therapy approach, we achieved robust chimerism when 6 × 10⁶ ET3-modified Sca-1⁺ HSPCs were transplanted, compared to low chimerism with 2 × 10⁶ ET3-modified cells.

We have previously shown engraftment outcomes can be impacted by inconsistencies in murine specific reagents, such as ATG, that are otherwise not present in current human transplant procedures. Russell et al. show the use of this multi-epitope antibody in a non-genotoxic conditioning regimen resulted in inconsistent engraftment because of binding to HSPC receptors, like CXCR4 that is fundamental for chemotaxis of stem cells to the BM.^29,54 Other preclinical studies also highlight that ATG-induced T cell depletion is not directly comparable between murine and human HSCT settings.^55,56 Therefore, in this study we included IdeS in our regimen to achieve enzymatic cleavage of ATG and ensure there is no residual ATG in circulation at the time of transplant that could interfere with receptors on transplanted HSPCs. In our anti-CD117-sap/ATG/5-FU regimen the IdeS is specific only to the rabbit ATG and not rat-generated antibodies, as this protease is specific for rabbit, human, monkey, and sheep IgG. We think this is a useful strategy for consideration in other preclinical conditioning regimens using antibodies featuring any of these subclasses. Hansa Biopharma’s clinical trials testing imflidase (IdeS) as a pretreatment have shown variable efficacy in gene therapy settings for Crigler-Najjar syndrome and Duchenne muscular dystrophy, where the patients are immune to adeno-associated virus (NCT06518005 and NCT06241950). However, IdeS may remain necessary for translation into a non-genotoxic conditioning regimen for autologous transplant of gene-modified HSPCs since clinical practice with ATG is currently limited to allogenic HSCT settings and preclinical data confirms it has a potential risk of impairing engraftment.⁵⁵ An alternative method for cleaving conditioning agents, which could be equally useful, is to target the neonatal fragment crystallizable (Fc) receptor with an inhibitor like efgartigimod.^65,66,67

In general, for HSCT, there is an overarching consensus that HSPC depletion is sufficient to achieve robust engraftment. However, this metric is not always applicable across all HSCT settings. For example, in Fanconi Anemia (FA), it has already been shown that HSPC depletion is not necessary since autologous gene-corrected or healthy donor HSPCs possess a competitive advantage over the failing FA HSPCs.^68,69 Furthermore, in WT mice we show even with an optimized ADC we do not achieve sufficiently deep depletion to permit consistent, robust engraftment of cultured gene-modified Sca-1⁺ HSPCs comparable to irradiation. Increasing the number of anti-CD117-sap doses administered could achieve deeper depletion, but in this study, we confirm ADCs like anti-CD117-sap can induce residual HSPC cycling. Therefore, increasing dosing could in turn exacerbate this proliferation effect and further impede engraftment of transplanted cells, especially genetically manipulated cells. Bankova et al. were the first to demonstrate antimetabolite repurposing in the context of HSCT conditioning, which tested 5-azacytidine in combination with an anti-CD117 mAb.²⁸ We confirm similar synergy between an ADC and an antimetabolite in the presence of immune suppression in which the addition of the antimetabolite enhances donor engraftment by targeting ADC-induced residual cycling HSPCs. However, we encountered engraftment issues when transplanting gene-modified cells, which was, at least in part, due to the formation of anti-ET3 IgG.

A potential limitation of this study is the pharmacology of 5-FU. Although metabolites of 5-FU can be DNA damaging, there is a growing body of literature showing the agent’s primary mechanism of action is actually through metabolites that induce RNA damage.³⁹ Therefore, if 5-FU is advanced, it will be important to demonstrate its mechanism of HSPC killing. Nonetheless, we think 5-FU serves as a proof of concept drug that highlights how antimetabolite repurposing can be successfully incorporated in ADC-based non-genotoxic conditioning regimens. Although we show conditioned mice retain weights and successful breeding post-transplant, our study shows limitations in assessing fertility since reproductive systems vary greatly between mice and humans, and thus these data are likely not directly translated to the human condition. Therefore, this too will require additional studies. Another limitation is our lack of understanding why mouse strains with higher WBCs and lower CFU counts showed increased survival and engraftment compared to those strains that showed the opposite. Other groups have also reported on prior strain differences, and we think these parameters could be biological indicators of regimen sensitivity since these values directly impact partitioning of the conditioning agents.^18,28,69 However, a potential confounder of the observed HA-E16 mortalities is that mouse strains display varying mast cell reactivities and CD117 expression levels on mast cells. Therefore, additional studies testing antihistamine treatment prior to anti-CD117-sap/5-FU/ATG conditioning are warranted. It should also be noted that our studies utilize an HSCT model relying on the transplantation of WT donor Sca-1⁺ cells into conditioned HA recipient mice to allow for tracking of engraftment levels, which are routinely used in murine gene therapy settings to represent CD34+ cells. It should be understood that this model is not fully representative of an autologous transplant setting, but our findings can inform optimization of conditioning platforms. For example, we have historically utilized this HSCT model in preclinical studies and shown it has allowed for translation into clinical settings.^8,11,29 Specifically, our insights on overcoming engraftment and immune barriers in non-genotoxic conditioning regimens can have translational relevance for autologous HSCTs currently relying on busulfan or treosulfan in combination with fludarabine to enhance immune suppression and facilitate engraftment of gene-modified cells. Lastly, follow-up studies are needed for statistical evaluation of these findings since comparisons were not performed due to the binary nature of the outcomes measured across some exploratory studies.

Although we have shown anti-CD117-sap with the immunosuppressive antibody cocktail permits engraftment of gene-modified cells, engraftment was variable and not comparable to that which we achieve with conventional irradiation. Therefore, in this study, we focused on elucidating the biology underlying inconsistent engraftment and incorporating a mechanism-based repurposing of antimetabolites into the regimen to augment engraftment levels. Collectively, we show non-genotoxic conditioning regimens are not universal and must be optimized to account for the subsurface variations that exist in these HSCT protocols, such as HSPC counts, baseline CBCs, disease phenotype, ADC biochemistry, and selection of other reagents to be included in the regimen. Our data also suggests measuring HSPC depletion by flow on the day of transplant is not a reliable indicator of successful engraftment outcomes. However, we found the combination of LSK-SLAM and HSPC-CFU assays can function as an engraftment screening predictor. This is especially important as a quality control measure for assessing conditioning regimens that feature biologics with inherent variability across batches. We also show that conditioning with ADCs like anti-CD117-sap can induce a proliferative state in residual stem cells that have the potential to outcompete incoming donor cells at the time of transplant. Other non-genotoxic conditioning regimens based on antibody targeting of HSPCs likely produce a similar cycling phenomenon that could result in low level engraftment. Therefore, repurposing FDA-approved antimetabolites in combination with HSPC-targeting ADCs may play an important role in augmenting HSCT engraftment with non-genotoxic conditioning regimens, especially in a gene therapy setting like HA.

Materials and methods

Mice

Exon 16-disrupted HA mice were obtained as F1 hybrids [S129 and B6] from L. W. Hoyer (Holland Laboratory).^60,70 These HA mice were either backcrossed 5 generations to create an E16 colony (mixed 30% S129 and 70% C57BL/6 (B6) backgrounds) or >10 generations onto a B6 background to create an MIRB colony (>97% B6 background). Male and female recipient mice 8–12 weeks of age from both the E16 and MIRB (CD45.2⁺) colonies were used in the in vivo BM depletion, cycling, and transplantation studies. Sex-matched B6 congenic mice 6 weeks of age from Jackson Laboratories (Stock #002014) carrying the differential Ptprc^a pan leukocyte marker (CD45.1⁺) or green fluorescent protein (GFP) mice from Jackson Laboratories (Stock #006567) were used as donors in all transplantation studies. In addition to being used as donor mice, the CD45.1 and GFP strains were used as recipients throughout the studies as WT comparisons for the results obtained from the hemophilic mice. All mouse strains were housed and bred at an Emory University Division of Animal Recourses facility. Animal studies were performed under a protocol approved by the Emory University Institutional Animal Care and Use Committee.

Protein expression and purification of immunotoxin

A small amount of a master glycerol stock of a verified Cysteine 255 Saporin-3 clone was inoculated into 2 flasks of 200 mL LB Broth supplemented with 100 μg/mL ampicillin and 34 μg/mL chloramphenicol.⁷¹ Cultures were grown overnight under 220 rpm shaking at 37°C in a New Brunswick Scientific incubator. The following morning, the overnight culture was distributed equally between 4 flasks of 1,000 mL LB Broth fortified with the same concentrations of antibiotics and incubation was continued under 230 rpm shaking at 37°C until the culture reached an OD_600nm of 0.8. Once reached, sap expression was induced with 1 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) for an additional 2 h under the same temperature and shaking conditions. The 4 flasks of 1L cultures were spun down and pellets were resuspended into a combined 200 mL PBS wash before final harvest at 3,400 rpm for 30 min at 4°C (Thermo Scientific Sorvall RC 6 Plus Refrigerated Centrifuge, FiberLite F14-6X250 LE Rotor). The pellets were frozen at −80°C for solubilization overnight and then thawed before a volume of 8 mL of BugBuster (EMD Millipore, Cat #70584-4) supplemented with 2 mM Dithiothreitol (DTT) and 5 mM Ethylenediaminetetraacetic acid (EDTA) was added per 1 g pellet. Resuspended cells were incubated at room temperature on a shaker until the solution cleared. The pellet was diluted 5-fold in 10 mM phosphate buffer at a pH of 6.5 and then spun at 16,000 rpm for 30 min at 4°C (Beckman Optima L-70 Ultracentrifuge, Beckman Type 70.1 Ti Ultracentrifuge Rotor 70,000). Supernatant was kept at −80°C for chromatography. Purification and polishing steps were achieved using a Bio-Rad BioLogic LP system and HiTrap CM FF 16/10 column, and combination of four supernatants prior to capturing provided purification yields of around 10 mg per single sap batch, as determined by the molar extinction coefficient and experimentally calculated via the Beer-Lambert Law [ε₂₈₀ = A/(c ⋅ L)]. After polishing, elutions were buffer exchanged into phosphate-buffered saline (PBS) using Amicon Ultra-15 Centrifugal filter Ultracel 10L filters.

ADC preparation

Preparation of anti-CD117-sap consisted of two distinct conjugation phases; detailed information pertaining the phases of conjugation is provided in Figure S1. During the first phase, 3 mg of polished sap was reduced with a final concentration of 0.7 mM TCEP on ice for 1 h before being buffer exchanged into Reaction Buffer 1 (40 mM sodium phosphate, 150 mM sodium chloride, 1mM EDTA) using Ultra-0.5 10K filters (UFC501024) to a final volume of 250 μL. Meanwhile, 3 mg of streptavidin (Prospec, Cat #Pro-283-b) was brought up to a concentration of 2 mg/mL in Reaction Buffer 1. After, a sulfosuccinimidyl 6-(3'-(2-pyridyldithio)propionamido)hexanoate) crosslinker (SPDP; Thermo Scientific, Cat #21650) was added at a final concentration of 0.5 mM, followed by incubation at room temperature for 1 h. The streptavidin-SPDP was also buffer-exchanged into Reaction Buffer 1 using the Ultra-0.5 10K filters to a final volume of 250 μL. The reduced sap and linked streptavidin-SPDP were then combined into a single mixture and incubated overnight at 4°C. The following day, the mixture was buffer exchanged into Reaction Buffer 2 (40 mM sodium phosphate, 150 mM sodium chloride) using first Ultra-0.5 50K filters (UFC505024) to remove free sap and then Ultra-0.5 100K filters (UFC510024) were used to remove the free streptavidin. During the second phase, the biotinylated anti-mouse CD117 monoclonal antibody (mAb) clone 2B8 (BioLegend, Cat #105804) was combined in a 1:1 molar ratio by vortex with the in-house-developed streptavidin-sap product.

Gel electrophoresis

Following each phase in the ADC preparation, a portion of the conjugation reaction product was subject to gel electrophoresis analysis under both non-reduced and TCEP reduced conditions. Samples were diluted in 5× Pierce Lane Marker Non-Reducing Sample Buffer (Thermo Scientific, Cat #39001) with or without 10× Bond-Breaker TCEP Solution (Thermo Scientific, Cat #77720). The samples were then boiled for 5 min at 95°C and ran for 30 min at 200 V on a Bio-Rad 4%–15% Mini Protean TGX Precast Protein Gel. Afterward, the gels were stained for 1 h with gentle shaking at room temperature in GelCode Blue Safe Protein Stain (Thermo Scientific, Cat #1860957) before being de-stained with molecular grade water overnight. Gels were imaged the following day on an Odyssey CLx using Image Studio (v.5.2) using the following settings of 84 μM, Medium, 1.0 mm.

In vivo administration of conditioning agents

In-house anti-CD117-sap immunotoxin was injected retro-orbitally as a single dose of 0.75 mg/kg on days −5 to −3 before HSCT, as specified per each individual experimental timeline. Rabbit anti-mouse antithymocyte globulin (ATG; Cedarlane Labs, CLAG3940T) was injected intraperitoneally (i.p.) at a dose of 30 mg/kg on days −5 or 3. A 10 mg/mL stock of 5-fluorouracil (5-FU; Sigma, Cat #F-6627) was prepared in sterile PBS by dissolving the 5-FU at 37°C and then measuring the absorbance at 265 nm. The 5-FU extinction coefficient (7.07 mM⁻¹cm⁻¹) was then used to calculate the true concentration for dosing. Twenty-four hours after anti-CD117-sap administration, the 5-FU was injected i.p. at doses ranging from 25 to 50 mg/kg, as specified per experiment. Fludarabine (NV Actavis, NDC. 45963-609-55) at a stock concentration of 25 mg/mL was injected i.p. at either a single dose of 200 mg/kg or given in three separate 30 mg/kg doses over the course of three days. IdeS (Acrobiosystems, Cat #IDS-S5143) was injected retro-orbitally as a 20 U/g dose in a total volume of 100 μL sterile PBS. Anti-CD4 (Bio X Cell, Clone GK1.5, Cat #BE0003) and anti-CD8 (Bio X Cell, Clone YTS169.4, Cat #BE0117) were injected i.p. at a dose of 200 μg in 100 μL of sterile PBS. Anti-CD40L (Bio X Cell, Clone MR-1, Cat #BE0017-1) was injected i.p. at a dose of 500 μg in 100 μL of sterile PBS. The detailed information and catalog numbers for conditioning reagents are also provided in Table S2.

Additionally, some conditioning experiments in this study note the addition of either FVIII prophylaxis or splenic leukocyte coverage. Twenty-four hours prior to immunotoxin administration, the mice in these experiments were injected retro-orbitally in a 100 μL total volume with either 2 μg of FVIII (KOVALTRY) or 2 × 10⁶ splenocytes from the same strain of mice. These splenic cells were isolated on the day of injection and subject to a single RBC lysis step for 10 min at room temperature using an in-house RBC lysis buffer (0.16 M Ammonium Chloride, 0.01 M Tris-HCl, [pH 7.5]). After lysis, cells were resuspended in sterile PBS.

BM depletion and cycling flow cytometry

Conditioned mice were sacrificed, and one dissected femur was flushed using a 23-gauge needle and the other femur was subject to crushing using a mortar and pestle. Flushed cells were counted on an automated cell counter after diluting 1:2 in the RBC lysis buffer mentioned earlier in the materials and methods section. Meanwhile, crushed BM cells were strained through a 40 μm cell strainer and then subject to 3 mL RBC lysis for 10 min at room temperature. All crushed BM cells were stained in PBS for 20 min in darkness, shaking at room temperature for flow cytometry detection of HSPCs using 1 μL of each of the following LSK-SLAM markers: lineage cocktail, CD117 (c-Kit), Sca-1, CD48, and CD150. After staining, cells were washed once with PBS and resuspended in 300 μL, prior to data acquisition on a Cytek Aurora analyzer. Dead cells were excluded from the analysis using 3 μL of 7-AAD Viability Dye (BioLegend, Cat #420404) that was added 2 min before the sample was run.

For the cycling analyses, femurs were processed as described previously but only 1.5 × 10⁷ crushed BM cells were processed for the LSK-SLAM staining. Stained cells were subject to a fixation/permeabilization step using eBioscience Foxp3/Transcription Factor Staining Buffer Set (Invitrogen, Cat. #00–5523) according to manufacturer’s specifications. Cells were then subject to intracellular staining with 7.5 μL Ki-67 for 45 min in darkness at room temperature. Cells were washed and resuspended in 150 μL PBS. Ten minutes prior to data acquisition on a Cytek Aurora analyzer, the BM cells were stained for 10 min in darkness with 1 μL of a 1:10 dilution of a 300 μM DAPI stock. Flow cytometry for both BM depletion and cycling experiments was performed using FlowJo software (v.10) and absolute counts were calculated by multiplying the percentage of the population of interest by all preceding gate percentages and the total flushed BM count. Detailed information and catalog numbers are provided in Table S2.

Hematopoietic CFU assay

For assessment of whole BM cell differentiation after conditioning, femurs from experimental mice were processed for flushing, as detailed earlier in the materials and methods section. After counting, 8 × 10⁴ BM cells were resuspended in 400 μL StemPro 34 media (Gibco, Cat #10640-019) and then added to 4 mL of Methocult GF M3434 (STEMCELL Technologies, Cat #03434). The suspension was vortexed and allowed to sit on ice for 10–15 min until the bubbles settled. Around 1 mL of the suspension was then plated on sterile 35 mm plates in triplicate. Plates were incubated at 37°C and 5% CO₂ for 10–12 days, as specified per experiment. Afterward, under a microscope, we counted the total number of colonies grown from erythroid progenitor cells (BFU-E), granulocyte-macrophage progenitor cells (CFU-GM, CFU-G, and CFU-M), and multi-potential granulocyte, erythroid, macrophage, megakaryocyte progenitor cells (CFU-GEMM). For assessment of differentiation after Sca-1⁺ HSPC transduction, 12 × 10³ Sca-1⁺ HSPCs were processed for plating and counting, as described earlier.

Pharmacokinetic clearance studies

To determine the level of residual circulating anti-CD117-sap, plasma was isolated from conditioned B6 mice on days 3, 4, 5, 6, and 7 following a dose of 0.75 mg/kg. Residual anti-CD117-sap (ng/mL) in 1:100 diluted plasma was detected per manufacturer specifications by an enzyme linked immunosorbent assay (ELISA) Rat IgG Kit (Innovative Research, Cat #IRTIGGKT) using a rat IgG standard. Absorbance at 450 nm was measured using a VersaMax Microplate Reader.

HSPC isolation, LV transduction, and transplantation

Murine Sca-1⁺ HSPCs were isolated from donor mice by immunomagnetic bead selection, as previously described.⁷² To briefly summarize, the BM was harvested by flushing as detailed earlier in the materials and methods section and strained through a 40 μm cell strainer before being incubated with a biotinylated Anti-Mouse Ly-6A/E antibody (Invitrogen, Cat #13-5981-85). Afterward, it was subject to a second incubation with anti-biotin microbeads (Miltenyi Biotec, Cat #130-090-485) and then the labeled cells were isolated using an LS column on a Miltenyi MidiMACS Separator. The isolated Sca-1⁺ HSPCs were then cultured at a concentration of 2 × 10⁶ cells/mL for 3 days in serum-free StemPro-34 SFM media (Gibco, Cat #10640-019). Culture media was prepared with a nutrient supplement (comes with StemPro-34 SFM media as a 40× solution), L-glutamine (0.29 mg/mL), penicillin-streptomycin (100 μg/mL), and the following cytokines purchased from R&D Systems: murine stem cell factor (mSCF; 100 ng/mL), murine interleukin-3 (mIL-3; 20 ng/mL), human interleukin-11 (hIL-11; 100 ng/mL), and human Fms-like tyrosine kinase 3 (hFlt-3; 100 ng/mL). On days 2 and 3 of culturing, the Sca-1⁺ HSPCs were transduced at a total MOI of 10–50 given in split vector doses. The second viral vector dose was given in the morning and then that afternoon the transduced cells were harvested, washed, and then resuspended in sterile saline for subsequent transplant by retro-orbital injection into preconditioned mice.

Peripheral blood collection and donor chimerism analysis

In the weeks following transplant, periodic retro-orbital bleeds were performed on all transplanted mice. The whole blood was collected into tubes with 3.8% w/v sodium citrate anticoagulant to prevent clotting and after each bleed a portion was aliquoted for CBC analyses on a Sysmex XN-1000V Hematology Analyzer. The remainder of the whole blood was centrifuged at 2,000× g for 15 min at 4°C to isolate and freeze plasma at −80°C for subsequent downstream analyses. After removal of the plasma, the remaining blood pellet was lysed using the in-house RBC lysis buffer detailed earlier in the methods. After three 10 min lysis steps, the cells were stained with eBioscience Fixable Viability Dye (Invitrogen, Cat #65-0865-14) and the following surface marker panel to assess donor versus host engraftment across the peripheral blood leukocyte types: CD45.2, CD45.1, CD3e, B220 (CD45R), and Ly-6G/Ly-6C. Following the staining a Cytek Aurora analyzer and FlowJo software (v.10) was used for data acquisition and analysis, respectively. Details of all antibodies used in chimerism flow cytometry are provided in a table in Table S2.

ET3 activity measurement

Following transplant with gene-modified cells, a Chromogenix Coatest SP Factor VIII chromogenic assay kit (Diapharma, Cat #K824086) was used per manufacturer specifications to assess ET3 activity levels in plasma collected from mice. Analyses were conducted using a standard curve generated with Factor Assay Control Plasma (George King Bio-Medical, Cat #0020-1) after measuring absorbance at 405 nm with a VersaMax Microplate Reader.

Anti-ET3 IgG ELISA assay

An ELISA assay was performed on mouse plasma to assess for the presence of anti-ET3 IgG post-transplant. Briefly, 96-well high-binding microtiter plates were coated overnight at 4°C with 25 μL of 1.5 μg/mL FVIII in coating buffer (20 mM Bicine, 2 mM CaCl₂, [pH 9.0]). The next day, the plates were washed twice with wash buffer (20 mM HEPES, 0.15 M NaCl, 2 mM CaCl₂, 0.05% Tween 20, 0.05% sodium azide [pH 7.4]) before being blocked for 2 h at room temperature with 160 μL of washing buffer plus 2.0% bovine serum albumin. The FVIII coated and blocked high-binding plates were then washed twice before adding 25 μL of 1:20 serially diluted mouse plasma and incubated at room temperature for 1 h. At the end of the incubation, 25 μL of 1:1,000 goat anti-mouse IgG-alkaline phosphatase (AP) in blocking buffer was added to each well. The plates were incubated at room temperature for 1 h and then washed twice. Afterward, 40 μL of AP substrate was added to each well and the reaction proceeded at room temperature for 20 min before being quenched with 0.4 M NaOH. The absorbance at 405 nm was measured with a VersaMax Microplate Reader and ELISA titers of each mouse sample were calculated using the Excel Titer Calculation Sheet.

VCN analysis

Proviral genomic DNA (gDNA) was first isolated from peripheral blood cells and transduced Sca-1⁺ HSPCs per manufacturer specifications of a DNeasy Blood & Tissue Kit (QIAGEN, Cat #69504). Afterward, VCN was determined in the gDNA samples by quantitative real-time PCR using Rev response element (RRE) primers specific for the sequence present in the ET3-LV constructs.⁸ Once isolated, 50 ng of each gDNA sample was analyzed in triplicate using 1× Power SYBR Green Master Mix containing 250 nM of the RRE primers. RRE plasmid DNA that was suspended in 50 ng of HEK293T/27 gDNA as background was serially diluted to create a standard curve for analysis on a QuantStudio 3 Real-Time PCR system.

Statistical analysis

Most outcomes including engraftment and transgene expression are binary under the non-genotoxic conditioning regimens tested. Apart from the pilot transplantation studies, a minimum sample size of five mice transplanted per group was determined to be sufficient to detect a minimum difference in group means of 50% assuming an alpha (α) value of 0.05 and power (β) of 0.80. Even though HA is an X-linked recessive disorder, we used both male and female mice per cohort to control for sex as a biological variable. Kaplan-Meier curves analyzed using the log-rank test were used for all survival analysis post conditioning.

Data and code availabiliity

All data supporting the findings of this study are available within the paper and its supplemental information files or from the corresponding authors upon request, without undue reservation.

Acknowledgments

This work was supported by funding from the Hemophilia of Georgia, the National Institutes of Health (NIH), and the Atlanta Network Training in KUH Scientific Research (ATLANTIS) grant number TL1-DK136047. All flow cytometry was supported by flow cytometers maintained by the Pediatrics/Winship Flow Cytometry Core of Winship Cancer Institute of Emory University, Children’s Healthcare of Atlanta, Winship Cancer Institute, Emory University and NIH/NCI under award number P30CA138292. All figures and schematics in this publication were created using BioRender.com.

Author contributions

J.O. designed research, performed research, collected data, analyzed data, and wrote the manuscript; J.S.A., S.R.P., G.M.B., and C.B.d.Q. performed research and analyzed data; W.H.B., C.P., and H.C.B. provided reagents; H.T.S., C.B.d.Q., and S.C. conceived and designed the research, analyzed data, and edited the manuscript. All authors reviewed the manuscript.

Declaration of interests

C.B.d.Q. is a cofounder of Expression Therapeutics, with equity, which owns intellectual property associated with ET3.