NK Cells: Null No More¹²

In preparing for this address, my first thought was, “How did I end up an immunologist?” As an undergraduate at Virginia Polytechnic Institute and State University (Virginia Tech) in 1975, Dr. Klaus Elgert, a new Assistant Professor in the Department of Biology, provided my first exposure to immunology. Klaus organized the first undergraduate course in immunology ever taught at Virginia Tech. I was a student in that first class, a class that Klaus still teaches today. Without Klaus, I probably would not be giving this address today, so thanks Klaus!

Infected with a curiosity about the immune system, I went on to become a graduate student at the University of North Carolina at Chapel Hill (1975–1978), joining the lab of Dr. Geoffrey Haughton for my thesis work on B cell tumors. There I characterized the “CH” series of mouse B cell lymphomas that many of you have used in your studies. Sadly, Geoff cannot be here today for me to thank him, but truly he taught me how to do science and to have fun at it. Geoff and his best friend Bernard Amos nearby at Duke were extremely inspirational teachers. My postdoctoral fellowship (1979–1981) was at the University of New Mexico School of Medicine in Albuquerque with Dr. Noel Warner. Noel introduced me to hybridomas and flow cytometry, the hot new technologies of the ‘70s and ‘80s that revolutionized immunology. When Noel moved to become Director of the Becton Dickinson Monoclonal Center (BD) in Mountain View, CA, he invited me to start my own lab there. Biotechnology was in its infancy in Silicon Valley, and it was a very exciting time. I had the opportunity to integrate molecular biology, flow cytometry, and mAb technology to understand how the immune system functions.

When I started my lab at BD in 1981, Noel encouraged me to explore the recently identified “natural killer” cells, a functionally defined subset that had been identified based on their ability to kill tumors without prior immunization. At the time, they were frequently referred to as “null cells” because they were thought to be devoid of any cell surface markers. Even at the time, I thought this very unlikely and quite a silly name. By good fortune, my first postdoctoral fellow, Joe Phillips, had made a hybridoma against human “large granular lymphocytes” when he was a graduate student in the lab of my good friend, George Babcock, then an Assistant Professor at M. D. Anderson Cancer Center. We discovered that Joe’s mAb, “Leu 11” (now designated CD16), recognized the Fc receptor on NK cells that is responsible for Ab-dependent cellular cytotoxicity activity (1). Our most exciting discovery at that time was that the subset of human PBLs expressing CD16 mediated the “natural killing” against the prototype K562 target cell. Moreover, by using two-color and three-color FACS analysis (then state-of-the-art) we found that the CD16⁺ lymphocytes did not express surface Ig or CD3, as shown in work published in The Journal of Immunology in 1983 (1). Thus, for the first time we had isolated a homogeneous population of NK cells and showed that they were distinct from B and T cells. This discovery, 25 years ago, started my quest to understand: How do NK cells develop? What receptors do NK cells use to recognize their targets? What is the role of NK cells in immunity against pathogens and tumors?

A real pleasure of science is the eureka moment (from the Greek heureka, “I have found it!”). I have been fortunate to share many “eurekas!” with my colleagues, fellows, and students in our pursuit of understanding the enigmatic NK cell. Today, I would like to share some of those moments with you.

Following the discovery that CD16 could be used to identify human NK cells, we wanted to know more about this receptor and about how it signaled. After a year of perplexing results we found that humans have two CD16 genes: one expressed exclusively by granulocytes and the other by NK cells, macrophages, and some T cells. A single amino acid difference in the proteins encoded by these genes causes CD16 on granulocytes to be anchored by a phosphatidylinositol linkage, whereas CD16 on NK cells has a transmembrane-spanning domain (2). However, we were unable to express the transmembrane-bearing CD16 protein until we realized that, like the TCR, CD16 requires the ζ-chain for stable expression and signaling (3).

In 1990, my research team moved to the DNAX Research Institute in Palo Alto, CA. DNAX was a wonderful place to do science with a fantastic group of colleagues. The pace of discovery and the excitement were remarkable. I only regret that my dear friend Jacques Chiller, President of DNAX during that time, is not with us tonight so I could thank him for his support and mentorship. While at DNAX, the eureka moments were too numerous to report in the next few minutes. I will share with you a few highlights from the decade in which Joe Phillips and I jointly directed the lab and comentored several postdoctoral fellows and research associates who have contributed to our understanding of NK cell biology.

At DNAX, our interest in NK cell recognition led to our discovery of a second ligand for CD28. Miyuki Azuma (a postdoctoral fellow in our lab), while studying the role of CD28 costimulation in the generation of human CTL (4), observed that the NK leukemia cell line YT expressed CD28 and was able to kill target cells expressing B7 (5). Miyuki perceptively noticed that the killing was completely blocked by an anti-CD28 mAb but was only partially blocked by an anti-B7 mAb. By screening for a mAb that completely blocked YT-mediated killing of B cell targets when combined with anti-B7.1, Miyuki identified B7.2. Together with another fellow in our lab, Chamo Somoza, we cloned the human CD86 gene in 1993 (6).

Meanwhile, in Stockholm Klas Kärre had observed that mouse NK cells preferentially kill certain tumors lacking MHC class I ( 7). This prompted us to ask whether human NK cells behave likewise. In collaborative work with Peter Parham at Stanford University, we conducted studies on HLA recognition by human NK cells. These efforts resulted in our discovery of the inhibitory NKB1 receptor, now designated KIR3DL1, which specifically recognizes HLA-Bw4. In 1994, Virginia Litwin, a postdoctoral fellow in our lab, made the prototypic DX9 mAb against KIR3DL1 (8). Another fellow, Annalisa D’Andrea, with help from a talented research associate, Chiwen Chang, cloned the gene encoding this receptor (9).

With the discovery of the killer cell immunoglobulin-like receptor (KIR)⁴ and Ly-49 receptors we were beginning to understand how MHC class I inhibited NK cells, but there was abundant evidence that NK cells do, in some circumstances, kill target cells that express abundant MHC class I. This led me to propose that the commonly held notion of “global inhibition” of NK cells by MHC class I was overly simplistic. It seemed more likely that NK cells were regulated by the integration of positive and negative signals. In 1997, to demonstrate this principle I performed some very simple “redirected” cytotoxicity assays using the Fc receptor-bearing mouse P815 target cells and human NK cell clones. When cytotoxicity was triggered by using mAbs binding to the activating receptors DNAM-1 (identified by Akira Shibuya, a postdoctoral fellow in our group), CD2, CD16, or CD69, NK cell-mediated killing was significantly blocked by coengaging the inhibitory KIR3DL1 receptor. However, when mAbs against these activating receptors were combined, ligating the inhibitory KIR3DL1 receptor was completely unable to prevent killing of the target (10). Thus, a balance of positive and negative signaling determines NK functional responses, a well-accepted concept today but a revelation 10 years ago.

After understanding (at least pretty well) inhibitory receptors, I have been driven by curiosity about the receptors and ligands that activate NK cells and how these provide immune defense. In 1997, one of my most exciting evenings in the lab was when I discovered DAP12 (11). Certain receptors in the KIR family lack ITIM but have a charged amino acid in their transmembrane regions, reminiscent of CD16 that we had worked on previously. After excluding the CD3 subunits and FcεRIγ, I suspected that another signaling adapter with a charged residue in its transmembrane segment likely existed. The observation of Vivier and colleagues supported the notion that a small protein coimmunoprecipitated with certain KIRs (12). By searching the DNAX database of dendritic cell-derived expressed sequence tags with the protein sequence of CD3δ, I found a small amount of sequence that appeared interesting. Further analysis revealed a small protein with a perfect ITAM and a transmembrane region with a centrally located aspartic acid residue. After cloning the full-length open reading frame, our lab team (comprising Brian Corliss, Clement Leong, Kathy Smith, and Jun Wu) confirmed that indeed DAP12 forms an activating receptor complex with the activating KIR, CD94/NKG2C, and Ly-49D receptors (13, 14). Soon thereafter, Lex Bakker (a postdoctoral fellow in our group) identified a macrophage receptor, MDL-1, associated with DAP12 (15). To date, >20 receptors expressed by NK cells, dendritic cells, macrophages, and granulocytes, as well as some T cells, have been shown to pair with DAP12.

Early one morning, shortly after we had discovered DAP12, Joe Phillips came into the lab and announced, “Lew, I’ve found another DAP!” DAP10 was born (16). By database searching, Joe had discovered a molecule with surprising similarity in its transmembrane region to DAP12. Remarkably, it was only a few base pairs away from DAP12 in the genome. Unlike DAP12, DAP10 did not have an ITAM, but we were excited by the presence of a YIMN sequence in its cytoplasmic region, a motif present in CD28 that recruits the p85 subunit of PI3K. We had thought that DAP10 might form complexes with DAP12; however, DAP10 was unable to associate with KIR2DS2 or CD94/NKG2C and we were unable to coimmunoprecipitate DAP12 with DAP10 (17). For some time we had been trying, without success, to express the orphan receptor NKG2D. Based on a hunch, I suggested to Jun Wu, a postdoctoral fellow in our lab, that he try cotransfecting DAP10 and human NKG2D. Again, “Eureka!”

The title of my Presidential Symposium is “From Academics to Biotech and Back,” which is appropriate because in 1999 I joined the Department of Microbiology and Immunology and the Cancer Research Institute at the University of California, San Francisco (UCSF), a wonderfully collaborative environment with outstanding immunologists. A few highlights of our research at UCSF include Heidi Cerwenka’s discovery and characterization of the ligands for mouse NKG2D, i.e., Rae-1 and H60 (work she started at DNAX and continued at UCSF) (18, 19). Melissa Lodoen (my first graduate student) demonstrated that mouse CMV (MCMV) encodes proteins that prevent the expression of the NKG2D ligands Rae-1 and H60 in MCMV-infected cells (20, 21). Hisashi Arase made the seminal observation that both inhibitory and activating Ly-49 receptors directly bind to the MCMV protein m157, suggesting that a virus is driving evolution of this family of receptors (22). Koetsu Ogasawara’s work implicated the NKG2D pathway not only in cancer and viral immunity but also in autoimmunity and transplant rejection (23, 24). Most recently, I was delighted and amazed by Jessica Hamerman’s discovery that DAP12 has inhibitory as well as activating function (25, 26).

Finally, I will share with you some recent, and yet unpublished, work from Joe Sun, who has been exploring the consequences on NK cell differentiation and function of exposure to the Ly-49H ligand m157 during NK cell development. These exciting new findings suggest that, as occurs with T cells, the Ags they encounter early in their development also influence NK cells.

Unfortunately, the time today is too short to tell you about all of the wonderful studies that have been done by our lab over the years. Nevertheless, based on the science done with my dear colleagues, fellows, students, and research associates for more than two decades, I proclaim, “NK cells: null no more!”

I thank my early teachers and mentors, Klaus Elgert, Geoff Haughton, and Noel Warner, for introducing me to the field of immunology and showing me the excitement of doing research at the highest level. The real joy of our profession is working with bright, imaginative, and energetic students, postdoctoral fellows, and research associates from around the world. Thank you all! A special thanks to Joe Phillips, with whom I shared a joint lab for 17 years and many eurekas! Finally, thanks to Jim Allison for his kind introduction to my address today and for 20-plus years of shared science and friendship.

The author has consulting or advisory arrangements with Schering Plough Biopharma, Protein Design Labs, Novo Nordisk, Shanghai Genomics, Avipep, Entelos, Ginkgo Biomedical Research Institute, and Symphony-Dynamo.