How Leukemia Cells Use Sugar-Coated Proteins to Evade Immune Detection

Researchers have identified a previously unrecognized mechanism that leukemia cells use to evade the immune system, according to a study published in the journal Science on April 9, 2026. Scientists from the Broad Institute, Mass General Brigham, and the Dana-Farber Cancer Institute found that leukemia cells utilize a sugar-coated protein to create a physical shield that prevents immune cells from attacking the cancer.

The protein, known as CD43, is located on the surface of leukemia cells. In these cancer cells, CD43 is heavily coated in sugar molecules—a process involving sialylation and O-linked glycosylation—which forms a dense barrier. This barrier physically blocks macrophages, a type of immune cell responsible for engulfing and destroying debris and pathogens, from reaching and consuming the leukemia cells.

The Role of CD43 as a Physical Barrier

The study specifically focused on acute myeloid leukemia (AML) cells. The researchers determined that the effectiveness of CD43 as a shield is due to the specific size and length of the portion of the protein that extends from the cell surface. This length allows the protein to act as a physical obstruction rather than relying solely on chemical signaling to deter the immune system.

Notably, the research found that this physical blocking effect did not depend on several known macrophage receptors that typically detect sialic acid sugars, specifically SIGLEC 1, SIGLEC 7, and SIGLEC 9. This suggests the protection is primarily a result of the protein’s physical structure and sugar coating rather than a specific receptor-based interaction.

Research Methodology and CRISPR Screening

To uncover this mechanism, the research team employed a genome-wide CRISPR knockout screen. This method allows scientists to turn off genes one by one to observe which ones are essential for a specific cellular behavior. In this instance, the screen was designed to identify which genes helped human AML leukemia cells avoid being eaten by human macrophages.

By growing human AML cells together with human macrophages, the team was able to pinpoint the genes most strongly associated with immune evasion. The results highlighted that genes involved in adding sugars to surface proteins were the primary drivers of the leukemia cells’ ability to hide from the immune system.

Contrasting CD43 and CD47

The study also provided new insights into CD47, a protein widely recognized as a don’t-eat-me signal used by various cancers to avoid phagocytosis. While CD47 is known to strongly reduce the ability of mouse macrophages to engulf cancer cells, this study found that it only weakly suppressed phagocytosis when human macrophages were used.

This discrepancy between species suggests that previous research relying on mouse models may have overstated the role of CD47 in driving immune evasion in human AML. The researchers noted that this finding could help explain why some macrophage-boosting treatments targeting CD47 have failed to improve patient outcomes.

Implications for Cancer Immunotherapy

The discovery of the CD43 sugar shield opens new possibilities for the development of cancer immunotherapies. By creating treatments that block CD43 or strip away its sugar coating, clinicians may be able to expose leukemia cells to the immune system, making them vulnerable to attack by macrophages.

Discovering this phenomenon opens up a new avenue for cancer immunotherapy. We think that the concept of cancers hiding behind sugar-coated proteins is not only important for leukemia, but likely other cancers as well.

Todd Golub, a Dana-Farber pediatric oncologist, director of the Broad Institute and co-senior author of the study

While the current research focused on AML, the authors suggest that other types of cancer may employ similar sugar-coated proteins to shield themselves from immune attack. Identifying these proteins across different cancer types could lead to a broader class of immunotherapies designed to break down these physical barriers.