Assistant Professor University of Illinois at Urbana Champaign, United States
Introduction:: Red blood cells (RBCs) have been an attractive engineering target for drug delivery, imaging, hemostasis, vaccination, and other applications due to their long lifespan, deep tissue accessibility, and crosstalk with the immune system. However, direct engineering of circulating RBCs in vivo remains a significant challenge. Here we report an unprecedented RBC tagging and targeting technology that enables metabolic labeling of RBCs with chemical tags for subsequent targeted conjugation of diagnostic and therapeutic agents via efficient click chemistries. We demonstrated that unnatural azido-sugars can metabolically label RBCs with azido groups via metabolic glycoengineering processes in vitro and in vivo, and further demonstrated that the azido-labeled RBCs can mediate targeted conjugation of dibenzocyclooctyne (DBCO)-bearing agents via the click reaction between azide and DBCO. The conjugated agents could retain in the bloodstream for weeks in mouse models, which can be utilized to prolong the circulation half-life of drugs in different disease contexts. We validated this technology by treating type 1 diabetic mouse model with metabolic glycoengineering of RBCs and following DBCO-insulin injection. This RBC tagging and targeting technology not only enables the development of potent therapies for treating diseases but will also provide a unique platform for studying RBC metabolisms and immunology.
Materials and Methods:: To demonstrate the possibility of metabolic glycoengineering for erythrocytes, we first studied whether azido-sugars can metabolically label murine erythroleukemia (MEL) cells, a cell line that can differentiate into erythrocytes in vitro. Three types of azido-monosaccharides, tetraacetyl-N-azidoacetylmannosamine (AAM), tetraacetyl-N-azidoacetylgalactosamine (AAGal), and tetraacetyl-N-azidoacetylglucosamine (AAGlu), were first synthesized and characterized. The kinetics of AAM, AAGal and AAGlu labeling for MEL cell line was studied. We further investigated the in vitro labeling of isolated RBCs with AAM and AAGal.
To see if the RBCs can also be labeled in vivo, we directly injected AAM intravenously and detected the azido group by using DBCO-Cy5 and Nucleic acid dye with flow cytometry. We also did confocal imaging and western blot to confirm the azido group on the RBCs cell surface. Additionally, three biological properties of label and unlabeled RBCs -aggregation, ATP level and reducing environment- were also further studied by experiments. Next, we studied if the azido tag can be used to conjugate DBCO-cargo in vivo. The mouse RBCs were labeled with AAM and 3 days later DBCO-cy5 was injected intravenously. The Cy5 molecule was detected by flow cytometry.
To test if the conjugation of molecule can have impact for disease control, we used Streptozotocin (STZ) to induce the diabetes in the mouse. Then DBCO-insulin was synthesized and injected to the mice. The blood sugar level was monitored to see the blood control and the body weight and plasm insulin level were measured to see the long-time effect for the insulin conjugation.
Results, Conclusions, and Discussions:: In the MEL labeling study, AAM shows the most efficient labeling compared with other two types of unnatural sugar. The isolated mature RBCs labeling study shows the successful labeling azido group after 24 hours and proves the possibility of RBCs metabolic labeling for RBCs. Those two in vitro labeling studies showed the best labeling performance of AAM.
In vivo labeling of RBCs shows stable labeling for 6 weeks and the biological properties of AAM labeled RBCs showed no difference compare with unlabeled RBCs.
In the in vivo labeling study, the azido tag on the RBCs surface can be detected as quick as 24 hours post the final injection and maintained up to 6 weeks without interfere the biological functions of RBCs. While the azido group on white blood cells disappear quickly in 3 days. In the in vivo cargo conjugation study, after injection of DBCO-cy5, the Cy5 signal in AAM RBCs is significantly higher compared with unlabeled RBCs as quick as in 3 hours. The Cy5 molecule on cell surface can maintain for 6 weeks, indicating the conjugation does not influence the life span of RBCs. In the diabetic mouse model, the AAM labeled diabetic mice combined with DBCO-insulin shows the best blood glucose control in the glucose tolerance test. Also, the body weight and serum insulin level in AAM labeling and DBCO-insulin group were also the highest compared with other groups. In general, we proved that RBCs can be engineered with azido tag via metabolic glycoengineering, and the azido tag can be used for the disease control by adding DBCO to the medicine. This generalizable technology can be further applied to many biomedical engineering field, such as immune modulation and imaging.
