Oncoimmunology Meets Organs-on-Chip

Abstract

Oncoimmunology represents a biomedical research discipline coined to study the roles of immune system in cancer progression with the aim of discovering novel strategies to arm it against the malignancy. Infiltration of immune cells within the tumor microenvironment is an early event that results in the establishment of a dynamic cross-talk. Here, immune cells sense antigenic cues to mount a specific anti-tumor response while cancer cells emanate inhibitory signals to dampen it. Animals models have led to giant steps in this research context, and several tools to investigate the effect of immune infiltration in the tumor microenvironment are currently available. However, the use of animals represents a challenge due to ethical issues and long duration of experiments. Organs-on-chip are innovative tools not only to study how cells derived from different organs interact with each other, but also to investigate on the crosstalk between immune cells and different types of cancer cells. In this review, we describe the state-of-the-art of microfluidics and the impact of OOC in the field of oncoimmunology underlining the importance of this system in the advancements on the complexity of tumor microenvironment.

Keywords: Cell on Chip; Oncoimmuno chip; Organ on Chip; cancer immunology; human on chip; microfluidic device; personalized medicine; tumor microenvironment.

FIGURE 1

Tracking cell perturbation changes on chip during melanoma cell-immune cell interactions and after immune cell migration. A single cell area variation analysis is shown in immunocompetent (A) versus non-immunocompetent cells (B). The area of a cancer cell is analyzed at fixed time intervals (e.g., 1 min) from an acquired time-lapse video, after the immune cell enters the cancer cell compartment and interacts (A) or fails to interact (B) with the malignant cell. Migrated immune cells do interact for longer time with cancer cells and this reflects in a minor frame-by-frame variation of the cell area. Hypothetically, this status is indicative of a dying/apoptotic cancer cell. This model underlines the migrate-to-interact strategy of an immune cell. The vertical brown/blue dotted lines within rectangles symbolically represent the complexity of the chip architecture and delimit the immune cell compartment from cancer cell compartment. The brown/blue dotted lines and points within graphs depict the area values for each frame referred to each time interval.

FIGURE 2

Schematic pipeline for TME reconstitution on an OncoImmuno chip. The first part of the workflow consists in a purification process of the main components of the TME, such as immune cells, tumor cells and normal skin cells plus stromal cells (if needed). When the design strategy of the microfluidic device is completed, the cells are then reconstituted inside the chip loading units. This reconstituted chip, resembling to an OOC, is then analyzed by a microscope platform. The TME can be derived from patients’ primary tumor or from an in vivo experimental tumor.

FIGURE 3

Application of the FBF tracking analysis to study immune cells versus cancer cells crosstalk. In this schematic model, the starting time-lapse recorded during the experiment execution is processed and all the variables associated to tracking analysis are extrapolated (Data computation). Specifically, when an immune cell is identified it will be tracked in each frame (Frame 1–6) to yield its associated trajectory (blue lines and points). This is done for each cell identified in the initial frame (Frame 1) of the video sequence. At the end of its walking, if the immune cell interacts with an adjacent cancer cell the duration of this interaction is computed and stored as a single interaction time value. In this example, this value is provided by the interval time difference between Frame 6 and Frame 9. The brown-dotted lines inside the microfluidic chip symbolize the compartment structures and ideally delimit the two immune and cancer cell compartments.

FIGURE 4

Schematic representation of most typical tracks of immune cells in relation to the presence of cancer cells during a time-lapse from an OOC experiment. During its walking toward the cancer cell, an immune cell can assume one of the indicated pattern motions. When the cancer cell exerts no influence on immune cells these can assume a track described by Pattern 1, namely a casual motion with no specific direction. This scenario is indicated by a very low directionality ratio compared to an ideal linear track (black dashed line), short mean step lengths, low speed, and high angular deviation of the immune cells. Conversely, if an immune cell moves forward by adopting the Track Pattern 4, it does successfully reach the opposite chip compartment and interacts with the neighboring cancer cell. This track is featured by an high directionality ratio (high linearity), step length and immune cells speed, and low angular deviation. In Track Pattern 2, immune cells pass through the opposite compartment but are unable to interact with the cancer cell. In Track Pattern 3 the scenario is similar to that described by Pattern 4, except for a low directionality ratio and then a weak migratory ability toward the cancer cell. Track Pattern 3 usually ends with a successful interaction between cancer and immune cell. The brown-dotted lines inside the microfluidic chip symbolize the compartment structures and ideally delimit the two immune and cancer cell compartments.

FIGURE 5

Schematic representation of the step length components in a tracking immune cell moving toward a cancer cell within a microfluidic chip. Each step length is calculated by returning the doublet of coordinates in starting (x₁, y₁) and finish (x₂, y₂) points. A single step length is depicted by the vector s, that can in turn be decomposed in the Δy and Δx units (representing the projection of the s vector to Y and X axes, respectively). Each step is associated to its relative angle deviation α conventionally referred to the X axis. Vertical dotted brown line symbolizes the structural components of the microfluidic device and ideally delimits the two immune and cancer compartments. Top right orange box displays the key formula associated to the step length and angle variation computations.

FIGURE 6

Role of GAN in immune cell trajectory prediction. Machine learning by GAN represents a practical manner to predict the whole track of a migrating immune cell. In this model, GAN predicts the migratory trajectory of immune cells (A, B, C, D, and E) and is helpful to foresee the behavior of immune cells when these encounter a tumor cell during their walking. GAN uses a Generator Network and a Discriminator Network to develop the predicted tracks from Real and Random tracks, extrapolated from each frame of the time-lapse. The Real tracks will then be processed into the Discriminator Network, whereas the Random tracks will constitute the input variables for the Generator Network, where they will be further processed and entered in the Discriminator Network. This Network exploits specific sigmoidal functions (Real, Fake) to distinguish the true predicted trajectories from fictitious tracks. The final result is constituted by all the predicted tracks associated to their input tracks originally entered. In this model, the computation of the predicted tracks P_B and P_C does allow to assert that the immune cells B and C will interact with the neighboring tumor cell with meaningful probability. Conversely, the predicted tracks P_A, P_D, and P_E indicate that Immune cells A, B and E will interact with the same cancer cell with a very low probability. These information are relevant to check potential immune cells to be selected for post-migration Interaction time computations. Brown-dotted vertical lines symbolize the structural components of the chip and ideally delimit the immune and cancer compartments.

FIGURE 7

Schematic representation of a dual condition chip to study the dynamics of apoptosis-induced anticancer drugs. Immune cells are loaded in the central chamber of the device. In parallel, cancer cells were loaded in presence of an extracellular matrix (e.g., Hydrogels, Collagen, or Matrigel) and specific drugs to be evaluated (Condition 1, Condition 2). Assuming that the Condition 1 is represented by untreated cancer cells and Condition 2 by tumor cells exposed to a cell death inducer (or a drug combo whose final action is cell death), these different on chip parallel treatment trigger a gradually divergent behavior of the cancer cells, reflecting in the final generation of two different TMEs (TME 1 and TME 2). TME 1 generates when the Condition 1 is a pro-tumoral drug or no treatment. In these cases, cancer cells will proliferate and a growing tumor will develop (TME 1). Concomitantly, an opposite effect due to tumor cells treated with anticancer agents occurs in the opposite microfluidic chamber (TME 2). The TME 2 is shaped by immune cells, that migrate under the effect of apoptotic factors released by cancer cells upon treatment with anticancer drugs. How immune cells shape the TME development is strictly dependent on how cancer cells are treated in the two side compartments. The initial key event of the TME 2 is the generation of apoptotic bodies as a consequence of tumor cell killing operated by immune cells. When cancer cells are exposed to a drug combination (Condition 2), the associated Condition 1 at the opposite side chamber can be a single drug of that combination or no exposure. Vertical dotted brown lines symbolically represent the structural units of the device.

FIGURE 8

Establishment of PDX-based OncoImmuno chip platform for melanoma patient’s personalized therapy. Melanoma lesions and blood samples are collected from enrolled patients. and tumors are then implanted in PDX mice. When the tumors reach an adequate size, they are excised and stored in a Biobank with the associated blood samples. In parallel, samples are assayed for immunohistochemistry, Hematoxylin/Eosin staining and Gene expression profiling. When a patient is recruited to initiate a therapy, there are two alternatives to assay the sensitivity of anticancer drugs: in vivo testing with re-engraftment of the stored melanoma lesion (requiring 14–20 days), and OOC-based analysis (requiring 48–72 h). Immune and cancer cells are obtained from the patient’s blood and tumor tissue, previously stored in the Biobank. Cells are then loaded in several types of customized microfluidic chips, with simple (OncoImmuno chip 1) or complex compartmentalization levels (OncoImmuno chip 2). Complex compartmentalization may require the use of other bio-components mimicking the TME matrix and the stroma (Matrigel, Hydrogel and collagen). In OncoImmuno chip 1, immune cells are left untreated whereas cancer cells are exposed on chip to a drug or left untreated (control condition). In OncoImmuno chip 2, Immune cells can be left untreated or treated with a single drug (Condition 1), whereas cancer cells on the opposite side chamber are exposed to a drug combo (Condition 2). Time-lapse is then started and the resulting video sequence is used for automated single cell tracking. Drug sensitivity is assayed by the extent of migration of immune cells (Immune cell tracking). To assess the vitality of melanoma cells, FBF area variation of these cells is evaluated and interaction times between cancer and immune cells are assessed (Post-tracking interaction analysis). For OncoImmuno chip 2 preferential migration of immune cells toward Condition 1 or Condition 2 is determined. Tracking profiles, interaction times and the associated kinematic variables will be compared to patient’s clinical parameters and exploited to start the cancer patient’s personalized therapy. Vertical brown dotted lines symbolically depict microfluidic device structural units and ideally delimit the immune and cancer compartments.

FIGURE 9

Modular view of a Human-on-Chip with emphasis on a hypothetical on chip immune system model. A Multicellular chip module is a unit containing all the key cellular components and extracellular factors derived from a specific human donor’s immune cells. In the Immune system these organs can be represented by five main cell populations (Stem cells, DC/Macrophages, Granulocytes, T cells, and NK cells). These cell subsets are properly loaded to form the five equivalent Multicellular chip modules (Immune cell compartments). Several Multicellular chip modules can generate an OOC Network. In this modular view, Immune system can be defined as an OOC Network composed by five Multicellular chip modules. Several OOC Networks compose a superior network of device-based organ systems, namely the Multi-organ module. The Human-on-Chip is composed of several serially connected Multi-organ modules. A Multi-organ module can also contain several subunits (Multi-organ sub-modules) to be employed for some isolated and targeted experiments (Sub-modular experiment) in which the sub-module is separated by its own Multi-organ module. In the case of the Immune system on chip, a Sub-modular experiment may be planned by a comparison of two connected OOC networks derived from two Multi-organ sub-modules (Immune system versus Normal skin or Skin cancer, respectively). Such an experiment can be aimed at comparing the immunosurveillance network activity to normal skin versus skin cancer. These experiments can be carried out with the help of AI algorithms and system biology approaches. Brown dots ideally represent the interconnections between the depicted units. The light-blue arrows evidence the passage to gradually superior levels of module complexity.

FIGURE 10

From the dawn of microfluidic devices to the future Human-on-chip for oncoimmunology. Time-line of microfluidic chip evolution from simple Lab-on-Chip (LC) and CC (Cell-on-Chip) devices to Organ-on-Chip (OOC) platforms integrated by AI algorithms for on chip oncoimmunology applications, toward the development of complex multi-organs on chip for in-depth investigations.