A key obstacle to medical research across a range of diseases is our lack of understanding of the immune system. Wayne Koff, a professor at Harvard University, described his decades of HIV research, "Slowly, over time, we began to discover that we knew a lot about HIV at the molecular level, but we didn't know anything about ourselves. The reason is that the immune system is very complex.
Many diseases involve the immune system overreacting, underresponding, or having a mistargeted response. Deepening our understanding of the immune system is essential for a better understanding and** a range of diseases. In Part 1 of this series, I shared the network that immune cells communicate and coordinate through the use of protein messengers called cytokines. These networks are complex and not fully mapped. Studying these networks is part of determining how and why the immune system responds. Here, I will share more research on immune cell-cytokine networks.
Immune cells coordinate like social networks
Immune cells and the cytokines they use to communicate can be analyzed using the same techniques as social networks. A network can be represented as a matrix, where each column and row represent a node, and the matrix values indicate the presence or strength of the connections between them. This type of representation is the basis of how the internet network is represented for search engines such as Google. Matrix math is everywhere! Jeremy Howard and I studied at the University of San Francisco and FASTMany of its applications are described in the computational linear algebra courses designed and taught by AI University.
*immprot samples immune cells from human blood and uses proteomics (large-scale analysis of proteins) to study immune communication networks. The researchers considered 28 cell types in a stable and activated state, as well as more than 10,000 proteins. IMMProt uses several classical mathematical techniques to analyze networks, including principal component analysis, supervised clustering, and lasso regression analysis.
Cytokines bind to receptors on the surface of the cells that receive them. Receptors for some cytokines are present on many different types of immune cells (these receptors are said to be "widely expressed"), while receptors for other cytokines may be expressed on very limited cell types. The researchers found that there are two types of cytokine communication modes: for a given cytokine, there are either many types of cellular sending and several types of receiving, or there are several types of cellular sending and many types of receiving. It is a targeted, asymmetric exchange of information.
The authors studied the amount of afferent and outgoing communication between different cell types and determined the degree of crosstalk between clusters of different cell types. In the image above, the outer circle lists the cell type, and the middle circle lists the receptor and cytokine gene names. This ** shows eosinophils and basophils (two cell types that have evolved primarily to address parasites) send the cytokine CCL5 3 to a variety of other cells that receive information using the receptor CCR3. The color of the line indicates the type of cell that sends the cytokine.
While many of the relationships discovered have been established in existing science**, they have also discovered new ones and experimentally validated two of them: one introverted and one extroverted. "The immune system exhibits almost infinite diversity in terms of cell types and activation states, and achieving integrity in these two dimensions is particularly challenging for human cells," the authors wrote.
AI will not replace other ways of learning about the world: desktop research remains crucial
AI is a powerful tool, but it is not a substitute for other ways of understanding the world, including qualitative research (as I wrote with Luisa Bartolo) and benchtop quantitative experiments. Machine learning and AI techniques are very good at learning patterns in existing data. However, they are limited by the type of data collected so far, and it remains critical to continue new laboratory experiments. For this reason, the recent Dictionary of Immunity** is a valuable contribution to the field. The researchers collected direct measurements of more than 1,400 cytokine-cell type pairings in in vivo experiments.
This study is significant because it takes into account all major immune cell types and all major cytokines (studies typically focus on only 5 immune cell types) and is conducted in vivo (not in culture). The researchers injected 86 different cytokines into individual mice and measured the response of 17 immune cells in the mice's lymph nodes. They use single-cell RNA sequencing, which is more straightforward than considering ligand and receptor expression association data. They published their findings in the form of an "immunity dictionary" and developed companion computer software. The software is then used to identify cytokine networks in tumors after checkpoint blockade**, a type of cancer that helps reactivate depleted immune cells.
A key finding is that the cytokine-induced response is highly cell-specific. Rarer types of immune cells express more cytokine types than more common cell types. A rare cell type, often not even covered in immunology courses, has been found to express the largest number of different cytokines, affecting almost all other cell types. This result suggests that rare immune cell types are essential for cell-to-cell communication.
Another important finding is that each immune cell type can be polarized into multiple states, depending on the combination of cytokines it receives. Two key researchers were amazed at this plasticity and sophistication of immune cell types, with even the most well-studied cytokines inducing more complex responses than previously expected.
Areas of ongoing research
This is an exciting area of research. Much remains to be done to more thoroughly understand the intricate ways in which immune cells coordinate with each other to respond to threats. The immune cell-cytokine network is a good illustration of the power of interdisciplinary work, as NLP, mathematics, and benchtop immunology studies all provide important insights into this question. And this is just one of several important questions in immunology where artificial intelligence is applied!