AUTHORS
Stephanie Sendker, MD, Shaobo Yang2, Alice Chen3, Fuguo Liu, PhD4, Eden Bobilev5, Anna Clara Bader6, Maily Nyugen7, Mubin Tarannum8, Jenny Giang Nguyen Ho9, Veronica W Hui10, Kate Marinchenv11, Castilleja Albert12, Michael J James13, Bartosz Grzywacz14, Remy Dulery, MD, PhD15, Andreia Maia, MS, PhD16, Jerome Ritz, MD17, Robert J Soiffer, MD18, John Koreth, MD,MBBS,PhD,DPhil19, Jacqueline S. Garcia, MD20, Mounica Vallurupalli, MD21, Catherine J. Wu, MD22, Roman Shapiro, MD23, Andrew A. Lane, MD, PhD24, Rizwan Romee, MD25
AFFILIATIONS
1 Dana-Farber Cancer Institute, Boston, United States
2 Dana-Farber Cancer Institute, Boston, United States
3 Dana-Farber Cancer Institute, Boston, United States
4 Dana-Farber Cancer Institute, Boston, United States
5 Dana-Farber Cancer Institute, Boston, United States
6 Dana-Farber Cancer Institute, Boston, United States
7 Dana-Farber Cancer Institute, Boston, United States
8 Dana-Farber Cancer Institute, Boston, United States
9 Dana-Farber Cancer Institute, Boston, United States
10 Dana-Farber Cancer Institute, Boston, United States
11 Dana-Farber Cancer Institute, Boston, United States
12 Dana-Farber Cancer Institute, Boston, United States
13 Harvard Medical School, Boston, United States
14 Beth Israel Deaconess Medical Center, Boston, United States
15 Dana-Farber Cancer Institute, Boston, United States
16 Dana-Farber Cancer Institute, Boston, United States
17 Dana-Farber Cancer Institute, Boston, United States
18 Dana-Farber Cancer Institute, Boston, United States
19 Dana-Farber Cancer Institute, Boston, United States
20 Dana-Farber Cancer Institute, Boston, United States
21 Dana-Farber Cancer Institute, Boston, United States
22 Dana-Farber Cancer Institute, Boston, United States
23 Dana-Farber Cancer Institute, Boston, United States
24 Dana-Farber Cancer Institute, Boston, United States
25 Dana-Farber Cancer Institute, Boston, United States
ABSTRACT
Background Extracellular adenosine is a major immunosuppressive metabolite in aggressive tumors including acute myeloid leukemia (AML). To combat this, we engineered tumor-homing non-pathogenic Escherichia coli (E. coli) to surface-display the enzyme adenosine deaminase (ADA) to achieve sustained enzymatic activity, localized adenosine degradation, and immunometabolic reprogramming of the leukemic environment.
Methods Using our established E. coli scaffold-surface display system we previously demonstrated as a novel form of immunotherapy (Yang et al, Nat. Biotechnol. 2025), here we engineered E. coli DH5α to display ADA under a rhamnose-inducible promoter. Surface expression and enzymatic adenosine-to-inosine conversion were validated by liquid chromatography mass-spectrometry (LC-MS) and enzymatic activity assays. Adenosine was quantified in fresh bone marrow aspirates from 15 AML patients. The effects of ADA-displaying bacteria on primary human NK cells under adenosine (10mM) exposure were assessed by spectral flow cytometry, cytotoxicity assays (1:1 E:T ratio), metabolic profiling (Agilent Seahorse XF), and bulk RNA sequencing. In-vivo safety, therapeutic efficacy, bone marrow trafficking, and immune remodeling were evaluated in the syngeneic WEHI3B AML model.
Results Adenosine levels significantly increased in AML bone marrow at diagnosis vs. remission (33.2 vs. 0.8 µM, p=0.02) and strongly correlated with marrow blast burden (R2=0.8, p=0.003). Adding adenosine to NK cells reduced expression of key immune-activating markers on NK cells (CD25 p=0.004, NKG2D, p=0.006) and T cells (CD25 p=0.001), decreased NK and T cell proliferation (Ki-67: NK cell p=0.01; T cell p=0.02) and suppressed cytotoxicity of NK cells 24-fold (p<0.001) and CD70-CAR T cells 10-fold (p=0.006) against NOMO1 and OCI-AML3 leukemic cells. To target adenosine, we successfully engineered E. coli DH5a to surface-display murine and human ADA using different scaffold constructs. Comparative LC-MS analysis confirmed functional ADA activity, with murine ADA showing the greatest adenosine-to-inosine conversion after 4h (log10_INO/ADO-ratio: 2.67, p=0.02), whereas the best-performing human ADA construct reached 0.85. Under adenosine pressure, ADA-displaying E. coli significantly enhanced CD70 CAR T cell-mediated killing of OCI-AML3 cells (23% vs. 40%, p<0.001) and NK cell–mediated killing of NOMO1 leukemia cells (1% vs. 42%, p<0.001) when compared to scaffold controls. ADA-displaying E. coli overcame adenosine-mediated NK and T cell dysfunction by restoring proliferation (Ki-67: NK cells, 18.9 vs. 41.7%, p<0.001; T cells, 21.7 vs. 40%, p=0.002) and activation (NK cells, CD25 13.8 vs. 64.4%, p<0.001; NKG2D 17.6 vs. 45.5%, p=0.002; T cells, CD25 42.1 vs. 55.9%, p=0.004; NKG2D 10 vs. 20.2%, p=0.007). While adenosine (10mM) impaired maximal oxidative respiration in NK (137 vs. 85pmolO₂/min, p=0.03) and T cells (193 vs. 20pmolO₂/min, p=0.02), ADA-displaying bacteria fully rescued oxidative phosphorylation by restoring spare respiratory capacity (NK cells, 54 vs. 107pmolO₂/min, p=0.03; T cells, 11 vs 148pmolO₂/min, p=0.01) promoting metabolic flexibility under adenosine stress. Consistently, transcriptomic profiling revealed substantial enrichment of the OXPHOS hallmark gene set in NK and T cells exposed to ADA presenting E. coli compared with scaffold controls (NES=2.19, FDR=1.3×10-9 and NES=1.31, FDR=3.9×10⁻2).
In WEHI3B syngeneic AML model, intravenously administered ADA-displaying E. coli rapidly localized to the bone marrow (3h) and accumulated in the liver (106CFU) by day 4, enhancing hepatic immune-cell infiltration, increasing CD45⁺ leukocytes (1% vs. 6%, p=0.04) and CD3⁺ T-cell abundance (2.8% vs. 6.7%, p=0.01). This was accompanied by a significantly prolonged survival (median survival, 22 vs 13days; p=0.002) in E. coli-ADA treated mice and notable reduction of tumor load particularly in the liver (%GFP+ WEHI3B-cells Δ−26, p=0.03) and spleen (Δ−25, p=0.03), which showed the highest leukemia infiltration. Treatment was well tolerated, with stable body weight and no evidence of CRS or systemic toxicity by plasma cytokine profiling.
Conclusion We establish ADA enzyme-displaying bacteria as a first-in-class platform for immunometabolic reprogramming of the leukemic environment, that overcomes adenosine-mediated stress and demonstrates the potential of enzyme-based therapeutic approaches laying the foundation for broader applications in cancer immunotherapy.
DISCLOSURES
Stephanie Sendker, MD: No relevant conflicts of interest to declare.