The Regulation of Immunity

Good afternoon, and thank you, Peter [Jensen], for the terrific introduction. As you all know, my talk is prerecorded, so I have no idea what Peter said. But I still thank you, Peter, and I look forward to getting together with you in the future. I want to offer a special thanks to Michele Hogan and the American Association of Immunologists (AAI) staff and the AAI Program Committee and its chair, Dave Masopust, for putting this meeting together, for creating all the features and events, and for working with all of you so that your presentations can be heard as well. I hope everybody will take advantage of the features and the presentations.

Each AAI President is asked to write a statement that appears in the AAI Newsletter at the beginning of their term. When I was thinking about mine, I was thinking about the great science being done in the field of immunology, from immune therapies to trying to understand the molecular basis of how immune cells were responding. Really, truly it was, and is, an amazing time in science.

But, we were in the summer of 2019; there was political uncertainty and deliberate political manipulation. Science was often discounted, climate change was not believed by many, there were those who were urging that we shouldn’t get vaccinated for anything, and it was a very difficult time to deal with all of these issues. It reminded me of the beginning of A Tale of Two Cities by Charles Dickens, which seemed, even though it was written more than a hundred years ago, to echo the feelings I was having: “It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness…”

The quote really reminded me of our times, but it also said what our mission should be at the AAI. It is our mission to represent science and represent the truth. And so, one of the things that we did that year was a lot of letter writing. This effort was spearheaded by the AAI public affairs office (Lauren Gross and her colleagues) and the AAI Committee on Public Affairs chaired by Ross Kedl. Letters were written to various government agencies and Congress, and these letters represented immunology as a field, supported truth, and supported I want to especially thank them for their efforts in those letters. I hope we made a difference with them, and I believe we did.

Another part of that year was worrying about whether we were going to host our annual meeting, IMMUNOLOGY2020, in Honolulu. Unfortunately, with the coronavirus disease 2019 pandemic beginning to rage in March, we canceled the meeting, as did all other organizations. I want to thank Drs. Jenny Ting and Michele Hogan for making the gracious suggestion to share the IMMUNOLOGY2021 spotlight and combine our two annual meetings.

From a young age, I was interested in DNA, and as I went through school, I wanted to know, “How does DNA work?” For my graduate work, I chose the laboratory of Dr. Richard Zitomer (1978) because his laboratory was incorporating recombinant DNA technology that had just been invented. It was also incredibly controversial at the time, and being in an exciting and controversial field seemed like a really good idea. Being that Richard was a yeast molecular biologist and geneticist, my initial task was to characterize a series of yeast mutants that had defects in expression of the cytochrome c protein to see if the mutations correlated with mRNA levels. I went through the process of learning how to do Northern blots. In those days, you had to create a paper filter that covalently bound the mRNA for the blots. It was a difficult procedure as the materials were labile, and so everything failed if you let it sit for any amount of time. Fortunately, we no longer do any of that. We published the work in the Journal of Biological Chemistry (1). This was my first paper.

Another part of my project was to determine the 5′ and 3′ ends of the iso-1 cytochrome c (CYC1) mRNA. To do this, we chose to purify the CYC1 mRNA itself and sequence it. The idea was to start with a giant vat of yeast cells, isolate total RNA, work my way down to the poly(A) message, hybridize the mRNA to a denatured DNA probe for CYC1 that I bound to a filter, elute off the mRNA as a purified messenger, and have it enriched to sequence. Sequencing an mRNA was something that didn’t happen every day and everywhere but did happen in the laboratory of Dr. Michael Smith at the University of British Columbia in Vancouver. Dr. Smith was a pioneer in synthetic DNA chemistry, and he worked with Dr. Shirley Gilham, a DNA chemist in his laboratory. Dr. Gilham had synthesized an oligonucleotide to the cytochrome c gene that I could use to basically extend to the 5′ end using reverse transcriptase and map the place where it began.

So, the plan was set. I would purify the mRNA, and I would go to Vancouver to the Smith laboratory. I worked closely with Dr. Gilham. She taught me how to sequence RNA using her oligo primers, reverse transcriptase, and dideoxy sequencing. I also learned to sequence DNA when I was there. We found the 5′ and 3′ ends of the message in that one trip, and I came back and began writing up the paper, which I got to coauthor with my advisor, Dr. Gilham, and Dr. Smith (2). Dr. Smith himself was later recognized as both a pioneer in DNA chemistry and for inventing site-directed mutagenesis, which won him the Nobel Prize in 1993—none of my work contributed to that prize. My experiences as a student were phenomenal and shaped my career and interests in biochemistry, genetics, and molecular biology.

After completing my Ph.D., I wanted to study gene regulation in higher eukaryotes and, in particular, in human cells. The problem was that there weren't a lot of genes that had been cloned and/or systems set up to study gene regulation in human cells. Dr. Jack Strominger’s laboratory had just published a paper describing the cloning of an MHC class I (MHC-I) gene (3). I applied and joined the laboratory in September 1982. The Strominger laboratory was busy cloning MHC-I and MHC-II genes, so I joined the fray with the objective of cloning an MHC-II β-chain gene. Over the next few months, I made a genomic library and, using a partial cDNA clone, I pulled out a genomic clone for HLA-DQB1, which at the time was called DC-3β. I sequenced the gene over the next year, all 8000 base pairs of it.

One thing that is hard to appreciate today is that I didn't send the DNA off to be sequenced. To sequence the gene, I performed hundreds of sequence reactions, incorporating both dideoxy and Maxim–Gilbert methods, to get through difficult reads. Perhaps the most difficult part was assembling both strands and determining where there were errors. After assembling it all, it had to be typed up into a figure, and the annotation was also done by hand with rub-on letters (4).

I also participated in the identification and sequencing of other MHC-II genes with Thomas Spies, Kiyotaka Okada, and Charles Auffray (5–7), but an important collaboration that we had was with visiting fellow Tucker Collins, who was with Jordan Pober’s group. Together with Alan Korman, a graduate student in the laboratory at the time, we worked with Tucker to show that MHC-II mRNA was inducible by IFN-γ (8, 9).

And so, now that I had a gene to study, I began to ask questions about how MHC-II genes were regulated. To do this, I made a series of deletions in the 5′ promoter region of the DC-3β gene. We found two regions that positively controlled IFN-γ induction. Conserved regulatory sequences were just previously identified as being conserved in a few MHC-II genes (10, 11), However, when we compared the sequences with the rest of the genes that we had identified in the laboratory, these sequences were conserved in all MHC-II genes in mice and humans (12) and were responsible for IFN-γ expression (13). They've since been subdivided into X1, X2, and the Y box regions.

With a gene system and a regulatory system in mind, I had the tools and preliminary data to go start my own laboratory. I was hired at Emory University in the Department of Microbiology & Immunology by the chair, Dr. John Spitznagel.

The field of MHC-II regulation was quite large at the time, and there were several key reagents that made further understanding this system possible. These reagents consisted of B cell lines created by Roberto Accolla and Don Pious: RJ2.2.5 (14) and 6.1.6 (15), respectively, that were deficient for transcription of all MHC-II genes. There were also B cell lines derived from individuals who had a SCID called bare lymphocyte syndrome in which all MHC-II genes were present, but expression was absent. Genetic complementation data suggested that the cell lines had mutations that functioned in trans (16, 17). Using several of these cell lines, we showed that the transcription factors deficient in the mutant lines functioned through the X box region (18).

The 1990s were focused on identifying the factors that bound the cis-regulatory elements. Walter Reith and Bernard Mach identified the RFX complex binding to the X1 box (19). Carlos Moreno in my laboratory showed that RFX was composed of three different proteins (20), and each of these was a different complementation group matching the BLS cell lines (16, 20). Carlos also identified CREB1 as the factor binding to the X2 box (21). NF-Y was the Y box factor, and it was identified by the Mathis and Benoist groups (22). A big breakthrough came in 1993 with the cloning of a non-DNA binding factor called CIITA by complementation cloning, and this experimentation was done by the Bernard Mach group and Viktor Steimle (23). CIITA was the key: it connected the IFN-γ induction observation of Collins with MHC-II expression as it was inducible by IFN-γ and was the factor absent in non–MHC-II–expressing cells (24, 25). James Riley in our group showed that CIITA was in fact a transcriptional activator, and that domains of the protein were able to direct transcription (26). The rest of the decade followed with the cloning of the genes for the rest of the BLS factors. Uma Nagarajan in our group cloned and identified RFXB (27), which turned out to be mutated in the largest number of patients with bare lymphocytes syndrome (complementation group B).

We and others shifted our focus to studying CIITA regulation and the roles of histone-modifying proteins in regulating the system. By the mid-2000s, we had a model for how MHC-II genes were regulated by the DNA binding proteins, the target sites, the binding of a CIITA, histone modifiers, recruitment of RNA polymerase, and ultimately transcription.

Another field also growing at this time was the field elucidating the role of chromatin architecture. Among the particular proteins being examined was CTCF, described by Gary Felsenfeld (28) as an enhancer blocking protein or an insulator. CTCF could block enhancers from crossing its barrier and activating a gene. We now know that this protein forms the basis of long-range chromatin interactions and loops, and it serves to insulate or isolate heterochromatin from euchromatin. It works with the cohesin complex, which is a large ring structure that has an ATP-dependent motor that can extrude DNA through the ring forming large loops and bringing enhancers in close contact with promoters. A nice description of this is in Li et al. (29), which shows structural interactions between parts of cohesin and CTCF.

Jorge Gomez and Parimal Majumder in the laboratory were looking at the time for novel MHC-II genes distal regulatory elements and found a series of CTCF binding sites. We were curious about how they played a role in MHC-II expression. Most interesting was the fact that each of the MHC-II α-β gene pairs was surrounded by CTCF binding sites as if CTCF binding sites had to be duplicated along with an α-β gene pair and were important for its expression (30). The non–MHC-II genes in the region were also separated from the MHC-II genes by CTCF binding sites. Using chromatin conformation capture (3C) assays, we showed that CTCF-CTCF binding site interactions occurred in the MHC-II locus, and that such interactions were distant dependent with a 250-kb maximum distance. We also showed that the promoters of MHC-II genes interacted with their neighboring CTCF sites, also in a distance-dependent manner, but this occurred only if the gene was active—in other words, if the gene had CIITA bound to its promoter (31–33). If we knocked down expression of either CTCF or a cohesion subunit using short hairpin RNAs, MHC-II expression was diminished (30, 34). The same was true in the mouse; CTCF sites surrounded α-β gene pairs. Interestingly, while comparing CTCF-CTCF interactions in mouse B cells and plasma cells (PCs), we saw that the interactions in PCs rearranged such that it appeared as though CTCF isolated the MHC-II region from the rest of the genome (35). This model made some sense because MHC-II is not expressed in PCs.

We drew lots of models of loops and interactions of CTCF and cohesion, and the idea here was that when the genes are expressed, they come together and interact, with perhaps all expressed MHC-II loci interacting together (36, 37). With other gene systems, it appeared that inside the loop is where genes are expressed, and outside the loop, perhaps they're not being expressed. So, depending on how each cell is assembling its chromatin structure, you get expression or not. We do not know whether the architecture is dynamic on a per-cell basis or whether it is set.

A number of years ago, super-enhancers were identified as regions that potentially controlled several genes and were important for overall gene regulation (38, 39). Super-enhancers have a number of key features: they were larger than 10 kb in length; had high levels of histone H3K27 acetylation; an active chromatin modification; were bound by chromatin-modifying factors like P300, which would place that mark; were enriched in motifs for transcription factors; and often had a large number of single-nucleotide repeats that were associated with disease. In one of the original papers, Hnisz et al. (38) identified the entire MHC-II region between HLA-DRB1 and HLA-DQA1 as a super-enhancer. This region included their promoters, a CTCF binding site called XL9 that we discovered earlier (32), and a new region of high histone acetylation, which in itself had features of a super-enhancer.

Notably, this new region had an extraordinarily high level of polymorphism. Further analysis suggested that this region is the most polymorphic region of the human genome. A large number of genome-wide association studies correlate polymorphisms in this region to numerous diseases. So, you have a region that's more polymorphic than the peptide binding domains of class II molecules in the middle of the MHC. To see if this new region, which we termed the DR/DQ-SE, possessed regulatory function, we used CRISPR-Cas9 to delete it from a B cell line (37). Deletion of DR/DQ-SE resulted in reduced expression of the HLA-DRB1-DQA1 and -DQB1 genes, but not the CIITA or HLA-DRA genes. We examined what the DR/DQ-SE region was interacting with by 3C. We were able to show that the super-enhancer not only interacted with each of the promoter regions of the above three genes, but that it also interacted with the XL9 CTCF binding site. Additional 3C assays showed that the super-enhancer interacted with the other neighboring CTCF sites on the distal sides of the HLA-DR and -DQ loci. When we deleted the DR/DQ-SE, all of the interactions were lost. This finding led us to a model in which we thought that the DR/DQ-SE super-enhancer was dictating the interactions between the CTCF binding sites, as well as MHC-II gene promoters. We proposed that this model may be the basis for forming a single transcriptional hub regulating the MHC-II locus (37), which would require further testing and refinement.

The large number of single-nucleotide polymorphisms in the locus suggests that they are likely to be inherited with the different alleles because it's typically a haplotype that is inherited. Because of their location in the super-enhancer, we posit that the polymorphisms may control the levels of MHC-II. This theory predicts that two types of polymorphism control our ability to respond to an immune challenge: those associated with the peptide binding groove of MHC-II molecules and those that might dictate the level of expression of MHC-II molecules. Together these could control selection, T cell repertoire, and responses to disease challenges.

When one starts a new laboratory, usually a short time into it, they start another project. Linda Gooding, a faculty member in our department at Emory, was interested in how TNF affected immune responses, and we began a project to look at how TNF regulated gene expression. We isolated TNF-induced genes from a cell line—and we found two genes to work on: MnSOD or Sod2 and Mcp1 (40). We showed that both used NF-κB as a transcription factor. For ∼15 y, we studied the mechanism by which these genes were regulated and how NF-κB functions to control gene expression (41).

In the mid-2000s, we collaborated with Rafi Ahmed to study how PD-1 was regulated. Rafi put together a team to work on PD-1 (Bruce Walker, Daniel Kaufman, Rafick Sekaly, James Riley, Michael Dustin, Arlene Sharp, and Gordon Freeman). For our project, we mapped out the transcription factors and the major regulators of PD-1 gene expression, which included NFATc1, BLIMP1, and STATs, and defined epigenetic mechanisms that regulate the gene (42–45). Our PD-1 project is still active (2021) in the laboratory. Our recent PD-1 paper that we published in The Journal of Immunology examines the mechanism by which epigenetic regulators can downregulate PD-1 following acute infection (46).

The second of two science stories I'm going to tell you about today focuses on the differentiation of B cells into PCs, and the transcriptional and epigenetic processes that control this process.

As you know, the naive B cell is a very quiescent cell. It has a few basic functions: express Igs on the surface to capture Ag, express TLRs to respond to inflammatory stimuli, and present Ags through MHC-II to be stimulated or activated by T cells. Once it receives an antigenic stimulus, B cells become activated and proliferate, and some differentiate to PCs. We wanted to understand the transcriptional and epigenetic events that occur as PCs form. PCs can produce 8000 Abs a second. Compared with B cells, they have a very different morphology, and they use a lot of different metabolic processes. This activity requires a large change in the transcriptional programming of the cells.

When Benjamin Barwick joined the laboratory as a graduate student, he wanted to connect cell division with the differentiation events of becoming a PC. Ben basically adapted an in vivo adoptive cell transfer system in which he took splenic B cells, labeled them with CellTrace Violet to track divisions, and adoptively transferred them into a recipient mouse. He challenged the mice with LPS to activate B cells robustly and analyzed splenic B cells and PCs 3 d after challenge. We initially used µMT hosts because we could get sufficient cell numbers to analyze at the transcriptional and epigenomic levels. There are several important takeaways from our analyses (47). The first takeaway is that nearly all of the B cells were activated and began dividing. Second, roughly eight divisions occur in cells before they differentiate into CD138⁺ cells. CD138 (syndecan-1) is our surrogate marker for PCs. If we do that same experiment in wild-type animals, that is, transfer wild-type B cells into wild-type mice, we can get some PC formation before division 8. But if we were to do this experiment in a MYD88-deficient host, we can see that the results mimic the pattern seen in the µMT mice. These findings suggest that the early contribution of the B cells is through additional extrinsic factors that might be accelerating the process (48). Last, Ben sorted the cells by division and carried out RNA sequencing (RNA-seq) and DNA methylation analyses on the sorted cells (47). The results showed that there was both a general increase in the mRNA content of the cells and an increase in specific gene expression patterns over the general increase with naive B cells expressing some 35,000 transcripts and the PCs formed expressing ∼175,000. DNA methylation analysis by division showed that there is a general loss in CpG methylation as the cells divide, with ∼10% of the genome becoming hypomethylated. This hypomethylation was not randomly spread across the genome. Rather, it occurred in clusters where there were many hypomethylated CpGs, and these clusters were often associated with transcription factor binding motifs important for PC gene expression. Thus, B cells go through at least eight divisions to form Ab-secreting cells (ASCs), and during these divisions there is a progressive change in epigenetic programming that correlates with changes in the cell’s transcriptome.

Christopher Scharer (49) in the laboratory carried this one step further and used transposase accessible chromatin with sequencing (ATAC-seq) to look at changes in accessible chromatin in cells undergoing proliferation and differentiation. At the early divisions, 1–5, there were few changes in regions becoming accessible, but at division 8 there was a substantial number of changes that occurred. When we mapped these regions back to transcription factor motifs, we saw that regions becoming more accessible contained motifs associated with PC-specific transcription factors, such as IRF4, and those that were getting closed were associated with initial B cell–specific fate transcription factors. Thus, the programming process of the epigenome accessibility is associated with defining regions accessible to transcription factors that drive cell fate. However, not all the cells were undergoing PC differentiation, and they all seemed to be staggered at different points (by division) in the process.

This finding led to questions about heterogeneity in the response, and whether there were specific cell fate decisions being made. To try to address that question, Chris Scharer and Dillon Patterson (48) in the laboratory used that same adoptive transfer experiment and subjected the transferred cells to single-cell RNA-seq using the 10X Genomics platform. Based on their transcriptomes, the cells could be grouped into eight distinct clusters in a t-distributed stochastic neighborhood embedding (tSNE) plot. To get an idea of where in the differentiation scheme the cells in each of the clusters were, we overlaid RNA-seq data that we previously analyzed from naive B cells, activated B cells, and ASCs, as well as the division-specific RNA-seq data. This technique allowed each cluster to be assigned to a position in the differentiation process from naive cells through two clusters of ASCs. With multiple clusters representing activated B cells and different divisions, we turned to pseudotime projection analysis (50) to understand the relationships between the clusters, predict the trajectory of each cell compared with its neighbor, and determine if there were branch points in the differentiation process.

Pseudotime analysis revealed two bifurcations in the trajectories of the cells. The first branch point occurred early at around divisions 2–3 and the other splitting the ASC into two clusters. One trajectory led to ASC fate and the other did not, even though the cells had divided multiple times. Comparing the transcriptomes of clusters from the early bifurcation, we found that only the ASC-destined branch was enriched for Myc target genes and genes in oxidative phosphorylation pathways. Myc target genes control proliferation responses, and oxidative phosphorylation is required for the B cell to become a PC. If you block oxidative phosphorylation, B cells will not become PCs (51). The non-ASC branch was enriched in inflammatory response genes and never downmodulated L-selectin (CD62L) expression, which gave us a marker to experimentally separate the branches. Cells isolated from each of the branches by CD62L and CD138 expression, as well as division 8, showed that only the ASC-destined branch could ultimately secrete Ab. The ASC-destined branch was also enriched for BATF target genes. IRF4 target genes were also enriched in the ASC-destined branch and not the others. When Chris and Dillon repeated the experiment in IRF4-knockout B cells, only the non-ASC branch gene sets appeared and the BATF transcriptional program was lost, suggesting that IRF4 and BATF drive the ASC-destined and not the inflammatory path (48).

Thus, in response to stimulation in vivo, B cells go through a coupled process linking transcriptional and epigenetic reprogramming with cell division. An early decision point is made in which BATF, IRF4, and MYC drive ASC formation. The bifurcation at the ASC stage (which I didn’t talk about) separates cells by autophagy and endoplasmic reticulum stress. On the non-ASC path, the cells maintain B cell fate. They still express CIITA and MHC-II gene products and genes associated with inflammation. Lastly, more recent experiments suggest that these same paths and decisions occur in responses to multiple Ags, including NP-Ficoll and following influenza infection.

I've been working with the AAI pretty intensely for the last 18 y. It began with a phone call from Robert Rich in early 2003, and he said, “Hey, Jerry, do you want to be a deputy editor? I'm going to be the Editor-in-Chief of The Journal of Immunology and it'll be great to have you on board.” And I said, “Okay, what do I have to do?” And he said, “Well, you just have to oversee a bunch of papers.” I forgot to ask him how many a “bunch” was. It turned out that there were eight deputy editors, and there were over 4000 new papers a year, plus revisions, and so forth. After Bob's term completed, I was approached by several people who said, “You should be the Editor-in-Chief,” and I said, “Okay.” I was chosen as Editor-in-Chief in 2008 by the AAI Publications Committee, and I got to work with Kaylene Kenyon, who was the publications director. She was absolutely amazing. She taught me everything I needed to know about how to be an Editor-in-Chief, and I hope I was a good one. Thank you, Bob and Kaylene!

After I finished my term, I thought I was done. But then, Michele Hogan was thinking, “You know, I can get Jerry to do anything…” So, she calls me up on the phone and says, “Would you run for Council?” And I said, “Okay.” I joined the Council that next year and then served as Vice President, President (hence the President’s Lecture), and then as Past President. The AAI staff have been a truly devoted group to the missions of the organization. Thank you, Michele, for that phone call. It was a great and rewarding experience.

I want to especially acknowledge the members of my laboratory, past and present, for their dedication to the work we were doing and for the science that they contributed. I hope they all had a really good time and learned a lot from being associated with our group. I mentioned only a few of you, but all have been part of my life, and I am extremely grateful.

I've had really terrific collaborators over the decades, both from within Emory and throughout the world. I currently am enjoying terrific collaborations studying B cells, PCs, and autoimmune diseases with Frances Eun-Hyung Lee, Ignacio Sanz, Christopher Scharer, Frances Lund, and Troy Randall. Thank you all very much for your colleagueship and for making science exciting.

I’d like to thank my funding sources from over the years, which have included the National Institutes of Health (National Institute of Allergy and Infectious Diseases, National Institute of General Medical Sciences, National Cancer Institute), National Science Foundation, Damon Runyon Cancer Fund, and The Michael J. Fox Foundation.

I need to thank my mentors. I've already talked about Richard and Jack, but to John Spitznagel, the chair who hired me, thank you for taking a chance on me. To Richard Compans, my next chair, thank you for keeping me at Emory and allowing me to do the things that I enjoy most. To my colleagues and friends, Charles Moran, Jr., Gordon Churchward, Dean Danner, Stephen Warren, Kirk Ziegler, and of course, Peter Jensen, I greatly appreciated your advice on my papers and grants and helping me out through the years. And to Susan Eckert, who was a coauthor with me when we decided to write a book and a whole bunch of articles on mentoring and how to be a successful scientist: thank you, Susan, for your mentorship.

And of course, I need to thank my wife, Valerie, my daughter, Diana, my sister, Tami, my parents, and my in-laws for your unwavering support and love.

I hope you all enjoy(ed) the meeting and I hope to see you all at IMMUNOLOGY2022 in Portland. Portland is great. It's sunny all the time. Be safe. Be well. Aloha.

The author has no financial conflicts of interest.