A Review Maps Neutrophil Migration, Tumor Contact and Killing in Microphysiological Systems

Why Neutrophil Behavior Needs More Than a Single Endpoint

Neutrophils can promote tumor invasion and immune suppression, but they can also kill cancer cells or support other immune populations. An Advanced Science review from the University of Texas at Dallas, UT Southwestern Medical Center and Stanford University examines how microphysiological systems can help resolve these context-dependent effects. Its focus is the use of human cells in three-dimensional cultures and microfluidic devices to study cancer interactions and evaluate therapeutic strategies.

The review argues against reducing tumor-associated neutrophils to a single pro-tumor or antitumor identity. Molecular studies describe multiple states, while a functional response depends on where cells migrate, whether they contact tumor cells, how long contact lasts and how neighboring cells respond. A tumor-size measurement alone may obscure these different processes.

Human-cell systems also address specific species differences. The review notes differences in circulating immune-cell proportions, chemokine pathways and Fc-receptor expression between humans and mice. These differences motivate complementary human experiments; they do not establish that every human culture predicts a clinical outcome. The value of a model depends on which biological interaction it represents and which measurement it supports.

Spheroids and Organoids Test Recruitment and Immune Interactions

The review describes spheroids as accessible three-dimensional aggregates that can recreate cell contacts, diffusion gradients and barriers to drug or immune-cell penetration. Patient-derived tumor organoids add tumor-specific genetic and phenotypic features, although neither format automatically preserves a complete immune and stromal compartment. Co-culture supplies selected missing components for a defined question.

Among the spheroid studies discussed, human neutrophils increased the growth of A549 lung cancer aggregates, and a CXCR2 antagonist reduced neutrophil infiltration. Other models showed how neutrophil extracellular traps, or NETs, can impede contact between tumor cells and T cells or natural killer cells. In those systems, DNase-mediated NET degradation restored aspects of immune-cell killing. These are distinct experimental mechanisms: reducing recruitment and removing an extracellular barrier intervene at different stages of the interaction.

The review also surveys engineered neutrophils tested in glioblastoma spheroids and patient-derived organoid studies linking tumor genotype to neutrophil survival or immune suppression. These examples come from distinct cited studies, with differences in tumor origin, immune-cell source and culture format.

This diversity matters for therapy development. A decrease in tumor growth could result from altered neutrophil recruitment, direct killing, reduced invasion or restored activity of another immune population. Experiments that measure more than one component can make those alternatives testable.

Chips Separate Migration, Contact and Vascular Events

Microfluidic devices can place tumor and immune cells in separate compartments connected by extracellular matrix, pores or vascular structures. They allow controlled gradients and flow, while optical access supports time-lapse imaging and single-cell tracking. The review organizes applications around recruitment into tumor tissue, interactions within the tumor microenvironment and encounters in the circulation.

For recruitment, it describes systems in which human neutrophils move toward tumor spheroids through matrix. In a pancreatic cancer model, CXCR2 inhibition changed migration speed, directionality and displacement. These measurements describe how cells reach a target; they are not interchangeable with a tumor-killing measurement.

Within tumor compartments, experiments quantified the frequency and duration of direct neutrophil–tumor contact as well as invasion and proliferation. One subtype-comparison study used HL-60-derived neutrophil-like cells. The antitumor-polarized cells showed greater contact and infiltration and reduced tumor invasion, while tumor cytotoxicity remained limited across all three subtypes.

Vascularized models address another phase of disease. The review describes studies of circulating tumor-cell clusters, endothelial-barrier disruption and extravasation. Neutrophils could promote these events in some settings, whereas a bone-metastatic niche model showed increased cancer-cell death after neutrophils entered the tissue. Together, the examples illustrate why the anatomical compartment and timing of the interaction affect its interpretation.

A Behavioral Research Agenda with Practical Validation Gaps

The authors introduce the neutrophil “behaviorome” as the set of measurable cell behaviors. They propose recording dynamic features such as migration, pausing, contact duration, infiltration and killing, then linking them with transcriptomic, proteomic or other molecular data. Such measurements could connect a molecular cell state with what that cell actually does in a tumor-like environment.

Some of the supporting imaging and behavior-guided transcriptomic approaches were developed for T cells or natural killer cells. Applying the same concepts to neutrophils is a proposed extension, not an already validated neutrophil outcome-prediction system. The review similarly identifies patient-specific drug screening and multi-organ assessment as development priorities. Autologous tumor and neutrophil samples could capture individual variation, while connected healthy-tissue compartments could help assess unintended effects alongside tumor responses.

Several obstacles remain concrete. Primary neutrophils are short-lived and sensitive to isolation-induced activation. Donor variation complicates reproducibility, and many published systems use healthy-donor cells or neutrophil-like HL-60 cells rather than matched patient neutrophils. Matrix composition, medium, flow, assay timing and imaging analysis can all affect the result. PDMS devices may also absorb hydrophobic compounds, altering the effective experimental exposure.

The review calls for standardized cell sources, isolation procedures, benchmark datasets and comparisons across laboratories. Its contribution is a map of what different human model formats can measure and where the evidence still depends on proposed extensions. For drug evaluation, that map favors an explicit chain from recruitment to interaction to tumor response, supported by the relevant controls and model-specific validation.