The Importance of Antibody Diversity

Part 1: Immunology: The Basics of Antibody Diversity

00:00:07;23 My name is Hidde Ploegh.
00:00:08;23 I'm an investigator at Boston Children's Hospital in the program of Cellular and Molecular Medicine.
00:00:15;20 I will present two talks today.
00:00:18;17 The first one provides you with a more general introduction to certain aspects of the immune system.
00:00:25;03 And in the second half of the talk, I'll speak about some evolutionary anomalies in the immune system
00:00:30;24 that we've been able to leverage into a new class of tools that I think will be
00:00:35;20 of more general interest.
00:00:36;27 So, let me begin by giving you an introduction of host defense.
00:00:43;23 We generally consider host defenses composed of three layers.
00:00:48;18 Mechanical and chemical defenses, depicted in this diagram, as line 1,
00:00:53;04 probably hold at bay 98% or more of viruses and pathogenic bacteria.
00:00:59;18 But because these organisms come equipped with special tricks to cir... circumvent these barriers,
00:01:05;11 we have a backup system.
00:01:08;08 This is the combination of innate and adaptive immunity.
00:01:11;23 Layer 2, innate immunity in this cartoon, should be considered the rapid-deployment forces
00:01:18;02 of the immune system.
00:01:19;27 They can distinguish between pathogenic entities such as bacteria and our own tissues,
00:01:26;11 but do so with a limited degree of specificity.
00:01:29;14 The nice thing is that they respond very quickly.
00:01:32;13 And so, should defenses of the mechanical and chemical nature fail, usually innate immunity
00:01:38;11 deals with the ensuing problem.
00:01:41;02 But given the sophistication of pathogens and the tricks they've evolved, some of these
00:01:46;21 require stronger measures.
00:01:49;12 And for that reason we have adaptive immunity kick in.
00:01:52;18 This is a time-consuming process, but it allows us to distinguish, truly with pinpoint precision,
00:01:59;01 between pathogenic microorganisms and our own tissues.
00:02:02;15 So, what I'll do in the next segment is to describe one particular aspect of
00:02:09;04 the adaptive immune system, because this will become relevant when we discuss,
00:02:12;24 in the second part of my presentation,
00:02:15;01 some of the unusual properties of antibodies made by other vertebrate species.
00:02:21;08 This is an amplification of the cartoon that I've just shown you, and it provides
00:02:25;27 a little bit more specificity.
00:02:28;13 You have the pathogens coming in.
00:02:31;09 They come equipped, as I've said, with enzymes that would allow one to
00:02:35;22 break down these mechanical defenses.
00:02:38;04 They can inactivate some of these chemical defenses.
00:02:40;22 And so when layer 1 fails, innate immunity kicks in.
00:02:44;17 And here we have a combination of cells -- such as macrophages and dendritic cells --
00:02:49;18 as well as molecules -- proteins of the complement cascade and hormone-like substances referred to as cytokines --
00:02:55;15 that collaborate to provide protection.
00:03:00;09 In turn, the output of the innate immune system synergizes with adaptive immunity.
00:03:06;16 And this layer of defense really becomes important when innate immunity fails.
00:03:10;21 So, the products elaborated in the course of an innate defense prime the pump,
00:03:16;04 so to speak, and facilitate the ensuing adaptive response.
00:03:20;08 This comprises types of lymphocytes that I'll discuss in a moment.
00:03:24;08 But it's really the synergy between innate and adaptive immunity that makes a key contribution
00:03:29;19 to host defense.
00:03:32;20 If we look at the kinetics with which these processes unfold, it recapitulates some of
00:03:37;20 the items that I've already spoken to you about.
00:03:40;25 Innate immunity consists of molecules such as type-1 interferons, natural killer cells...
00:03:47;08 and these kick in literally within hours to days of exposure to the pathogen.
00:03:54;00 If we look at what happens to the virus titer -- if we deal with, say, an influenza virus infection --
00:03:59;12 we see that innate immunity can rapidly reduce the number of circulating virus particles,
00:04:04;27 albeit not to zero.
00:04:06;23 And it is at this point that adaptive immunity must kick in.
00:04:10;09 We have virus specific CTLs; the abbreviation stands for cytotoxic T lymphocytes.
00:04:17;00 And we have antibody titers that rise as the infection is being resolved.
00:04:23;09 In a first exposure, the rise in antibody titers is relatively modest.
00:04:30;09 And in a phenomenon referred to as immunological memory or recall response, the secondary exposure
00:04:36;13 rapidly leads to massive induction of both antibody titers, we have memory killer T cells
00:04:44;00 that kick in, and it's the combined action, again, of these antibodies and T cells
00:04:50;21 that manages to control the infection.
00:04:54;27 If we think of where these processes occur in the human body, we must consider the circulatory system,
00:05:01;07 which includes arteries and veins.
00:05:03;25 It's the high arterial pressure that allows some fluids to leave the bloodstream,
00:05:09;02 which must be returned to the circulation in the form of lymph.
00:05:13;00 This lymphatic fluid is filtered through specialized structures called lymph nodes.
00:05:18;02 And it's really in these lymph nodes that the immune responses of the adaptive type
00:05:21;24 take place.
00:05:23;06 We should consider the circulatory system as a means of trafficking.
00:05:28;03 It's the vehicle via which lymphocytes, from their site of origin, arrive at their final destination.
00:05:34;22 And so, by monitoring what happens in the bloodstream, we can only get a transient snapshot
00:05:39;24 of what a real immune... immune response looks like.
00:05:42;14 So importantly, all of the important events that start an adaptive immune response
00:05:49;11 take place in specialized lymphoid structures called lymph nodes.
00:05:56;22 On the right, you see the organization of the lymphatic structures in a human.
00:06:02;14 The little ball-like structures are the lymph nodes, through which lymph fluid is filtered.
00:06:07;07 And it's really in these specialized structures that adaptive immunity is initiated.
00:06:14;10 One important cell type that we will revisit later on in this presentation are
00:06:20;19 so-called dendritic cells, thus named because they have spines that very much resemble what one finds
00:06:25;13 on neurons.
00:06:27;06 And these dendritic cells are positioned throughout the body.
00:06:31;06 They are really the first point of encounter of a foreign invader with the immune system.
00:06:36;18 And it's the ability of dendritic cells to assess the presence of an invader,
00:06:42;04 to then process that information, and present it to the appropriate cell types within the immune system
00:06:46;20 that is responsible for proper orchestration of these immune responses.
00:06:52;28 If we ask, what cell types contribute to adaptive immunity?
00:06:57;02 They are really the lymphocytes that I'll speak about most.
00:07:01;08 If we consider the origin of lymphocytes, they all derive from a stem cell that arises
00:07:05;24 in the bone marrow.
00:07:06;24 These so-called hematopoietic stem cells give rise to all bloodborne cells,
00:07:11;20 including platelets, red blood cells, and so forth, as shown on the left branch of this slide.
00:07:17;02 But importantly, for the remainder of the discussion, will consider mostly the lymphocytes.
00:07:21;26 They originate from a common lymphoid precursor and, through a series of carefully orchestrated
00:07:27;09 differentiation steps, they give rise to both B lymphocytes and T lymphocytes, thus named
00:07:32;20 because of their bone marrow and thymic origin, respectively.
00:07:37;04 The output of B lymphocytes are so-called antibodies or immunoglobulins,
00:07:40;23 a diagram of which is shown in the top.
00:07:42;24 And I'll return in some detail to the structural features of this class of molecule.
00:07:48;17 But let me point out that these antibody molecules, or immunoglobulins, exert a number of functions
00:07:53;27 that can contribute to protection.
00:07:57;02 First of all, they enhance phagocytosis.
00:08:00;15 This is the process by which the dendritic cells that I've just mentioned can
00:08:03;27 acquire particulate matter and process it to cells of the immune system.
00:08:09;04 Antibodies can also assist the function of elements of the innate immune system.
00:08:13;17 On the bottom left, I've shown natural killer cells.
00:08:16;13 They can bind immunoglobulins through receptors specific for them.
00:08:20;14 And once their union has occurred, they can assist in the killing of targets to which
00:08:24;28 the antibody is bound.
00:08:27;25 On the top right, you see yet another mechanism by which antibodies can confer protection.
00:08:33;07 And this is complement-mediated cytotoxicity.
00:08:36;24 In addition to immunoglobulins that circulate in the bloodstream, there's a class of proteins
00:08:41;25 called the complement proteins that, when properly activated, can directly exert
00:08:46;19 a cytolytic effect, either on bacteria or, as shown in this particular example, on tumor cells.
00:08:54;22 And then finally -- and this is one of the earliest discoveries as far as immunoglobulin
00:09:00;08 function is concerned -- immunoglobulins or antibodies can neutralize bacterial toxins.
00:09:06;28 They can bind to virus particles.
00:09:08;24 And by covering the surface of these structures, render them pretty much innocuous.
00:09:13;17 So, these are the many functions of immunoglobulins.
00:09:16;27 And the one that I've left out so far is the one on the top left.
00:09:21;05 We also have practical applications of immunoglobulins.
00:09:24;15 And spectacular recent examples include the immunotherapy of cancer.
00:09:28;21 And, as I'll show in the second half of my talk, we can make derivatives of these antibody fragments
00:09:34;04 and use them for purposes such as imaging of immune responses, non-invasively.
00:09:40;12 So, what about the structure of immunoglobulins?
00:09:43;02 As this cartoon illustrates, they are proteins abundantly present in serum.
00:09:48;11 They're glycoproteins composed of two identical heavy chains, in dark blue,
00:09:52;27 and two identical light chains, in light blue.
00:09:55;26 The heavy chains are glycosylated, and the light chains and heavy chains are held together
00:09:59;24 by disulfide bonds.
00:10:02;00 Biochemists would like to shrink the immunoglobulin molecules into units that retain the capacity
00:10:06;26 to bind antigen.
00:10:08;08 And for this purpose, proteolytic digestion has been used.
00:10:12;03 On the bottom left, you see the products that result from digestion with the protease papain.
00:10:16;13 It results in the release of fragments that are so-called Fab fragments.
00:10:21;24 They are monovalent and retain the capacity to bind antigen.
00:10:25;28 If you wish to retain the capacity of bivalent binding,
00:10:29;26 an intrinsic property of the immunoglobulin molecule, pepsin digestion may be used.
00:10:34;11 And this allows the two antigen-binding fragments to remain linked through disulfide bonds,
00:10:39;27 as indicated on the bottom right.
00:10:42;25 If we look at the diversity of immunoglobulins as they occur in the typical mammalian species,
00:10:48;24 there is massive diversity in structure and function.
00:10:52;00 I won't have the time to discuss all of these diverse functions, but I do want to highlight
00:10:56;09 a few of the salient structural differences.
00:10:59;08 We have here this massive pentameric structure of a class called immunoglobulin M or IgM.
00:11:06;03 We have a version of immunoglobulins that's found in secretions such as tear fluid,
00:11:10;10 held together by an unusual protein called the J chain.
00:11:13;28 We have the IgE molecule, implicated in allergic reactions.
00:11:19;05 And what most of you are probably familiar with are the immunoglobulins of the IgG classes,
00:11:24;11 of which several subclasses exist.
00:11:27;24 Now, when we look at the ability of an antibody molecule to bind a foreign substance,
00:11:34;08 also called an antigen, we realize that the immunoglobulin contacts the antigen
00:11:39;02 at the very tip of this Y-shaped structure.
00:11:42;15 And because structural biologists have been able to solve the three-dimensional structure
00:11:46;11 of antibody fragments in complex with antigen, we know at atomic resolution exactly how these
00:11:52;17 acts of binding occur.
00:11:54;02 So, in this box here, you see at higher magnification the typical mode of interaction of an immunoglobulin
00:12:01;07 with its antigen.
00:12:03;00 You'll realize that the immunoglobulin, composed of two identical heavy chains
00:12:07;06 and two identical light chains,
00:12:09;00 uses elements of both to achieve this specific recognition.
00:12:12;22 So, in light blue, the variable region of the light chain;
00:12:16;08 in dark blue, the variable region of the heavy chain.
00:12:18;25 And it is through the tips of these very subunits that interactions occur with the antigen.
00:12:25;24 These include hydrophobic interactions, salt bridges, van der Waals interactions...
00:12:30;23 a perfectly complementary surface is created to confer specificity.
00:12:35;22 And we know that antibodies can achieve a degree of specificity that allows them
00:12:39;17 to distinguish between molecules that differ in as little as one proton.
00:12:44;04 The presence or absence of a hydrogen atom can make all the difference
00:12:47;20 -- whether or not an antibody recognizes its target or not.
00:12:52;12 So, if we consider the ability of the immune system to mount an immune response against
00:12:57;27 pretty... any... pretty much any foreign substance we throw at it, we must ask the question,
00:13:03;06 how does the immune system achieve this remarkable result?
00:13:07;09 So, first of all, biochemists, without recourse to any molecular genetic tools,
00:13:16;28 accumulated large numbers of sequences of immunoglobulin proteins.
00:13:21;04 And this allowed them to relate the primary structure
00:13:24;09 -- that is to say, the amino acid sequence of the immunoglobulin variable regions --
00:13:29;16 to their antigen-binding properties.
00:13:30;24 And by aligning multiple sequences of either the heavy chin or light chain variable regions,
00:13:36;13 several salient features emerged.
00:13:39;03 The so-called hypervariable regions, indicated in red, are precisely those regions in the molecule
00:13:45;23 that contact the antigen.
00:13:47;27 And if one compares a large number of different sequences, that is also where
00:13:52;04 the majority of sequence diversity is concentrated.
00:13:56;02 This is not to say that other residues cannot vary, as is clear from the gray bars,
00:14:00;23 which indicate the variability index -- the extent to which different variable regions might
00:14:05;06 differ from one another -- but the bulk of the variation occurs in these three hypervariable regions,
00:14:11;19 also called complementarity-determining regions because that is exactly where
00:14:16;28 the binding of the antigen occurs.
00:14:19;11 Now, if one were to consider a million different antigens against which we would like to
00:14:25;08 raise an antibody, and you calculate the amount of genetic information required to encode
00:14:30;19 that information in the germline of an organism, you quickly reach the conclusion that
00:14:35;17 you run out of sequence space.
00:14:37;17 There is simply not enough DNA to encode, at the DNA level, the structure of a million
00:14:44;15 distinct antibody fragments.
00:14:46;21 And this is a question that has puzzled immunologists for decades until, in the '70s,
00:14:51;23 the molecular mechanisms by which diversity is generated became to be understood.
00:14:58;02 It turns out that immunoglobulin genes are, like many eukaryotic genes, genes in pieces.
00:15:05;18 But there's an additional element of surprise, here.
00:15:08;24 In fact, when we create a functional immunoglobulin gene, it's not just about introns and exons
00:15:14;18 that require splicing to create a functional messenger RNA.
00:15:18;22 The very cells that produce these immunoglobulins reshuffle their genetic information.
00:15:23;12 This is called somatic gene rearrangement, and it accounts for much of
00:15:27;20 the diversity of the immunoglobulins as proteins.
00:15:31;17 On the top of this diagram, you'll see our current understanding of how the light chain locus operates.
00:15:38;08 In mice and humans, there are two types of light chains called kappa and lambda,
00:15:42;12 and I'll confine myself to a quick description of what happens for the kappa light chain.
00:15:47;14 We have a battery of variable region sequences, separated by intervening DNA, followed by
00:15:53;22 so-called joining segments, and, at some distance downstream of it, the remainder of
00:15:59;16 the kappa light chain, the so-called constant region.
00:16:03;01 In the course of B cell development, somatic gene rearrangements occur, and this allows
00:16:07;28 juxtaposition of a randomly chosen V gene element with a randomly chose chosen J segment.
00:16:14;15 And it's not until this rearrangement process is complete that we arrive at
00:16:17;28 a functional light chain.
00:16:20;15 You'll notice that I've indicated the presence of an enhancer.
00:16:23;23 The promoters that drive expression of a functionally rearranged heavy chain do not come
00:16:29;00 within controlling distance of these enhancers unless and until somatic gene rearrangement has occurred.
00:16:34;11 So, the rearrangement process achieves two things.
00:16:37;10 First, it creates a functional unit that can be transcribed and translated into what
00:16:42;21 we know to be a light chain.
00:16:44;19 And second, its expression, its transcription, is controlled by an enhancer,
00:16:49;14 the function of which requires the rearrangement process.
00:16:52;24 For the immunoglobulin heavy chain locus, the situation is somewhat more complex.
00:16:58;04 In addition to this battery of these V segments and J elements, we have interposed
00:17:03;22 a battery of so-called diversity elements.
00:17:06;15 And in this case, the rearrangement process makes use of V, D, and J rearrangement
00:17:11;27 to arrive at a functional heavy chain variable region.
00:17:16;24 There is, again, an enhancer, the reach of which does not extend into those V genes
00:17:23;09 that have yet to rearrange.
00:17:25;04 And it's only upon completement... completion of the rearrangement process that the VDJ combination
00:17:30;18 is placed within controlling distance of this enhancer to enable expression of
00:17:35;27 a functional heavy chain.
00:17:39;11 This process is perhaps best compared to the one-armed bandit.
00:17:43;04 Think of V, D, and J elements as three independently spinning wheels on a slot machine.
00:17:49;17 The B cell, in the course of development, pulls the handle, and some random combination
00:17:54;08 of these Vs, Ds, and Js emerges.
00:17:57;24 This is not the whole story.
00:18:00;15 In this particular diagram, I've recapitulated what I've just told you -- for the heavy chain locus,
00:18:06;24 a battery of these Vs, Ds, and Js.
00:18:09;17 And in the course of B cell development, these rearrangements to which I referred occur
00:18:14;12 in highly ordered fashion.
00:18:16;14 First we have the D-to-J rearrangement.
00:18:19;04 And what I've indicated here by this little segment of rainbow-colored material in between
00:18:23;19 is a phenomenon called junctional imprecision.
00:18:27;11 When a D and a J element are juxtaposed, the act of recombination itself produces
00:18:33;05 some imprecision at the joint, adding and subtracting nucleotides in an unpredictable fashion.
00:18:40;04 And as you might imagine, if you disrupt the reading frame, you have what is called
00:18:44;08 a non-productive rearrangement.
00:18:46;15 If you add multiple nucleotides, you can affect the primary structure of the final product.
00:18:52;26 And so this imprecision in the course of V, D, and J rearrangement contributes to
00:18:59;02 diversity of the final product.
00:19:01;19 Not only do we see this junctional imprecision when Ds and Js rearrange, it also applies
00:19:06;23 when Vs are brought in to hook up with the newly generated DJ combination.
00:19:13;07 And if that weren't enough, there is an enzyme called terminal deoxynucleotidyl transferase
00:19:18;15 or TdT.
00:19:20;09 And this enzyme, in template-independent fashion, adds random nucleotides whenever Ds and Js,
00:19:26;21 or Vs and Ds, are joined together.
00:19:29;22 This massively expands the diversity of the final product.
00:19:33;16 And so if we consider the problem of antibody diversity, it is the combination of
00:19:38;07 a random choice of Vs, Ds, and Js, but that information is strictly germline encoded.
00:19:42;21 But the very act of somatic recombination introduces an element of imprecision
00:19:48;04 whenever joining occurs.
00:19:49;21 And this allows massive expansion of diversity of the immunoglobulin variable regions.
00:19:55;00 So, this slide summarizes much of what I've told you already.
00:19:59;10 In this case, for the light chain, I've indicated the positions of variability.
00:20:05;02 On the bottom, you see these hypervariable regions to which I made reference.
00:20:09;09 And the constant region, as the name suggests, is invariant in sequence and doesn't
00:20:14;00 make contact with antigen.
00:20:15;25 It serves to mediate interactions between the various building blocks of the immunoglobulin
00:20:21;01 molecule itself.
00:20:22;11 These ovals are referred to as immunoglobulin domains, and they all share a conserved sequence.
00:20:30;16 If we consider the different manifestations of immunoglobulins as they occur on the
00:20:36;11 surface of a B cell, we realize that there's an important cell biological distinction to be made.
00:20:42;02 B cells make both membrane-bound immunoglobulin, and that very same immunoglobulin can be secreted
00:20:47;13 as well.
00:20:48;20 This is a process that's controlled by alternative polyadenylation.
00:20:52;24 Depending on which poly-A addition site is used, the B cell either produces
00:20:57;25 the secreted version or the membrane-bound version of that one-and-the-same immunoglobulin.
00:21:04;03 This foreshadows the important role of the B cell receptor in perceiving antigen and
00:21:08;10 allowing B cells to expand, but also to allow that very same B cell to release immunoglobulins
00:21:13;08 into the bloodstream, where they can exert their effect, for example, by neutralizing
00:21:18;04 a virus.
00:21:20;05 The B cell receptor also plays a key role in orchestrating the processes that I've just summarized.
00:21:25;15 So, in the absence of a functional heavy chain rearrangement, B cells fail to complete development.
00:21:30;28 The discrete developmental stages are characterized by the presence of so-called surrogate light chains,
00:21:35;24 in this diagram depicted as VpreB and lambda-5.
00:21:40;00 And only when those subunits all come together and form a properly assembled pre-B cell receptor
00:21:46;28 does the B cell enable rearrangement of the missing piece, which is the light chain.
00:21:51;17 So, this pre-B cell receptor, depicted on the left, is a necessary condition for B cells
00:21:57;19 to engage light chain rearrangement.
00:21:59;27 And it's only when all these processes are executed perfectly that we arrive at
00:22:05;04 a fully assembled B cell receptor at the surface of a B lymphocyte.
00:22:10;04 You'll notice these little red and yellow stubs.
00:22:12;26 These are coreceptors, referred to as Ig-alpha and Ig-beta.
00:22:17;01 And they're absolutely crucial, because the B cell receptor itself
00:22:20;06 -- the immunoglobulin subunits --
00:22:22;22 lack the cytoplasmic tails required for signal transduction.
00:22:26;10 It's the non-covalent association with these accessory subunits -- Ig-alpha and Ig-beta --
00:22:32;01 that allow so-called immunoreceptor tyrosine-based activation motif, or ITAMs,
00:22:38;18 cytoplasmically disposed, to recruit the requisite kinases that initiate internalization,
00:22:43;19 proliferation of B cells that properly engage the antigen, and so forth.
00:22:47;19 So, to summarize, this would be the structure of a B cell receptor as you would find it
00:22:52;19 on the typical resting B lymphocyte.
00:22:54;23 A membrane-bound version of the IgM molecule in non-covalent association with these
00:23:00;23 accessory subunits, Ig-alpha and Ig-beta.
00:23:03;04 And it's through these accessory subunits that B cell receptors fulfill most of their functions.
00:23:09;22 There's an added layer of complexity.
00:23:11;18 And we'll have to use that when we discuss, in the second part, the unusual attributes
00:23:16;14 of certain antibody molecules made by camelid species, and this is a phenomenon
00:23:20;27 referred to as class switch recombination.
00:23:23;05 Recall that at the outset I referred to the different classes of immunoglobulins --
00:23:28;03 the hugely complex pentameric IgM all the way down to the more simple IgG molecules.
00:23:34;19 It turns out that a given VDJ combination can be put in juxtaposition with the information
00:23:41;05 that provides the IgM molecule, the so-called new chains.
00:23:45;19 And by a process called class switch recombination, that rearranged VDJ cassette can be placed
00:23:51;06 upstream of whatever constant region you might require to execute the necessary functions.
00:23:57;25 This class switch recombination requires the involvement of the other major class of lymphocytes,
00:24:02;16 this... the T lymphocytes or T helper cells.
00:24:06;00 And there are accessory molecules such as the cytokine, IL-4, and enzymatic functions,
00:24:11;11 activation-induced deaminase expressed in the B lymphocyte, that are an absolute prerequisite
00:24:15;20 to execute the class switch recombination.
00:24:18;20 So at the end of the day, you might end up with an IgG-producing B lymphocyte which takes
00:24:24;11 this VDJ cassette and places it in juxtaposition, in my example, with the gamma-2 constant region.
00:24:32;22 In yet another example, you might take that very same VDJ combination and instead
00:24:37;10 hook it up to the alpha constant region, so that you may suit... that so that you may produce
00:24:42;11 this secreted version of the IgA molecule.
00:24:46;00 Now, how... how is all of this arranged?
00:24:50;17 It turns out that we have a detailed molecular understanding of how this somatic rearrangement process,
00:24:56;01 as well as the class switch recombination, occurs.
00:25:00;02 And unlike the enzymes involved in putting together V, D, and J elements, class switch recombination
00:25:06;00 requires the activity of activation-induced deaminase, expressed in B cells only when
00:25:11;28 properly contacted by T helper cells.
00:25:15;20 In a looping-out reaction, the rearranged VDJ combination is put in juxtaposition
00:25:22;09 with whatever constant region the B cell demands at that point in time.
00:25:26;11 And by physical excision of the intervening DNA, we may now connect the functionally rearranged
00:25:32;17 VDJ combination to whatever constant region we require.
00:25:37;00 Now, importantly, I refer to the role of helper T cells to execute this reaction.
00:25:46;11 To understand a little bit more about how these T cells operate, let me give you
00:25:51;06 the following information.
00:25:54;09 The professional antigen-presenting cells -- think of the dendritic cells which I showed at the very outset --
00:26:00;04 may acquire antigen, a foreign substance, by a process called phagocytosis.
00:26:05;14 Once the phagocytosed antigen has been internalized and delivered to the appropriate endocytic compartments,
00:26:11;17 these antigens are attacked by proteolytic enzymes and converted
00:26:16;11 into short peptide fragments that will be displayed on the surface of the so-called antigen-presenting cell.
00:26:23;09 There's a special class of molecules involved in this process.
00:26:26;14 These are the products encoded by the major histocompatibility complex,
00:26:30;23 to which I'll return as well.
00:26:32;20 And it's really the combination of these unique peptide- MHC combinations that will be recognized
00:26:38;03 by T lymphocytes by means of antigen-specific receptors.
00:26:44;01 The B cell is a specialized case.
00:26:46;21 It too can bind to antigen by virtue of the fact that expresses, at its surface,
00:26:53;06 the B cell receptor for antigen.
00:26:55;12 The B cell receptor for antigen is really the high-affinity capture device that
00:26:59;28 allows the B lymphocyte to probe what's in the external environment and bind only those protein antigens,
00:27:06;01 or other foreign substances, for which it is specific.
00:27:09;26 It does so by virtue of what we call an epitope.
00:27:13;25 This is a structural feature of the antigen itself that can be seen by the B cell receptor.
00:27:19;06 Now, B cells can internalize the B cell receptor when complexed with antigen.
00:27:24;00 And by the same mechanism that I've just described, proteolytic activity will chop up the foreign protein
00:27:29;28 into short synthetic fragments, which are bound by these MHC products and presented
00:27:35;18 on the surface of the B lymphocyte.
00:27:38;04 It is the T cell that now recognizes, by means of its antigen-specific receptor, the unique
00:27:44;19 combination of peptides derived from the original antigen, presented by products of the MHC.
00:27:50;26 And the key concept to understand here is that the features of structure that
00:27:55;25 allowed the B cell to recognize antigen in the first place may well be distinct from the fragments
00:28:00;22 generated from that antigen and presented via MHC molecules to T lymphocytes.
00:28:06;13 This phenomenon is called linked recognition, and it ensures that only those B cells that
00:28:11;11 have acquired antigen and present peptides derived from it to appropriately specific T cells
00:28:16;22 that an antibody response can ensue.
00:28:20;01 So, to integrate all of this, and without going through the details... on the far left,
00:28:25;19 you'll see dendritic cells acquiring antigen and presenting it to T helper cells.
00:28:30;13 In the right half, you'll see B cells acquiring antigen and presenting peptides to T cells
00:28:35;04 of appropriate specificity.
00:28:36;17 And when all is said and done, we have a productive interaction between the T helper cell,
00:28:41;27 which is antigen specific, and the B cell, that is antigen specific.
00:28:46;25 And so this is how we can orchestrate an immune response.
00:28:50;22 I mentioned the fact that there are two major classes of lymphocytes: the B lymphocytes,
00:28:55;19 which we just discussed, and T lymphocytes, which as we saw provide necessary help
00:29:01;17 and also generate so-called killer T cells, or cytotoxic T cells.
00:29:06;19 They have antigen receptors very much like the B cell receptors we discussed.
00:29:11;09 And they make use of very similar rearrangement processes, in fact employing the exact same
00:29:16;20 enzymatic machinery.
00:29:18;02 So, the T cell receptor, like its immunoglobulin counterpart, is composed of two subunits:
00:29:22;23 alpha and beta subunits.
00:29:25;04 And they, like their immunoglobulin counterparts, make use of V-to-J and V-to-DJ rearrangements,
00:29:32;08 as diagrammed in this cartoon.
00:29:33;28 Each element is flanked by the appropriate recognition signal sequences,
00:29:38;01 features of structure that are shared with the immunoglobulin variable regions
00:29:42;06 of the heavy and the light chain.
00:29:44;18 Now, T cells, as I've said, recognize antigen not in solution but bound to the products
00:29:50;28 of the major histocompatibility complex.
00:29:53;24 As diagrammed in this cartoon, you see a T cell receptor with its two subunits engaging
00:29:59;26 a class-I MHC product, thus named because it spans the lipid... lipid bilayer only once.
00:30:06;14 And these MHC products present these short snippets of foreign protein to antigen-specific receptors
00:30:12;15 on T cells.
00:30:14;03 In the second part, I'll have a few words to say about these so-called co-stimulatory
00:30:18;16 or checkpoints.
00:30:20;02 These are molecules that can fine-tune immune responses, and either enhance or inhibit
00:30:24;20 immune recognition by T lymphocytes.
00:30:27;00 Now, the MHC products are unique in structure because, notwithstanding the fact that
00:30:33;25 they are of unique and fixed sequence, they can nonetheless bind a vast diversity of peptides
00:30:41;12 by virtue of the fact that the architecture of the peptide binding pocket is designed
00:30:45;15 such that many peptides of different sequence can fit into one-and-the same peptide binding pocket.
00:30:52;14 The overall global structure of a class-I MHC product is composed of a heavy chain
00:30:58;18 in non-covalent association with its light chain, beta-2 microglobulin.
00:31:03;00 And it's this assembly that creates the peptide binding pocket -- this is the top view of
00:31:07;04 the very same molecule shown here -- into which peptides bind for presentation
00:31:12;20 to these antigen-specific receptors.
00:31:17;12 The way in which this system functions is that T cells are test-driven on MHC products
00:31:23;00 that present peptides from our own self proteins, which you ideally would like to ignore.
00:31:28;13 And it's not until a stressful situation such as cancer or infection occurs that
00:31:33;08 new peptides derived either from pathogen-specific proteins or tumor-specific antigens
00:31:39;18 make their appearance.
00:31:40;18 So, the immune system is taught to ignore peptides of our own proteins.
00:31:47;03 And what remains at the end of the day is a repertoire of T lymphocytes uniquely capable
00:31:51;18 of recognizing peptide-MHC complexes that differ from our own self-MHC products.
00:32:00;13 If you think of an infectious situation, in the absence of any immune recognition,
00:32:08;20 unopposed infection might result in the organism’s death.
00:32:11;23 We have lytic infections.
00:32:13;07 We have massive virus production.
00:32:15;19 And it is for this reason that we have components of the adaptive immune system, to fight specifically
00:32:21;01 these kinds of events.
00:32:22;14 I've mentioned the fact that antibodies can neutralize virus particles in the circulation.
00:32:27;05 That is one means of protection.
00:32:29;14 I've indicated the existence of so-called killer T cells, the CD8-bearing T lymphocytes.
00:32:36;02 CD8 is a glycoprotein marker uniquely confined to these killers.
00:32:41;06 And by means of their antigen-specific receptors, they recognize class-I MHC products that present,
00:32:46;18 for example, viral peptides as in this example.
00:32:50;20 But because many pathogens have replication times vastly shorter than the host,
00:32:57;02 they can acquire mutations that allow them to elude immune attack.
00:33:00;17 And that's depicted by the transition of this somewhat innocuous pink virus to the nasty red.
00:33:07;09 Many of these viruses do so by, for example, altering expression of class-I MHC products,
00:33:12;26 and that also happens to be one of the mechanisms by which cancerous cells can evade detection
00:33:18;02 by T lymphocytes.
00:33:19;23 If you eradicate expression of class-I MHC products, you're essentially invisible
00:33:25;22 to the cytotoxic T lymphocyte, and that gives you the upper hand in terms of virus production
00:33:31;11 or, in the case of a cancerous cell, replication.
00:33:34;11 Now, we know a great deal about the molecular details by which the class-I proteins acquire
00:33:40;00 their peptide cargo.
00:33:41;27 From a cell biological perspective, this is a very unusual and interesting series of reactions.
00:33:47;01 And it focuses on the function of the ubiquitin pathway.
00:33:51;06 Proteins in the cytoplasm are modified by ubiquitin in an enzymatic cascade that involves
00:33:56;06 these three classes of enzymes: E1s, E2s, and E3s.
00:34:00;15 And having modified our protein with multiple ubiquitin molecules, now these proteins
00:34:05;20 are poised for recognition by the proteasome, which in a highly processive fashion
00:34:10;16 destroys these proteins and produces peptides capable of being recognized by T lymphocytes.
00:34:15;23 The problem, however, is the fact that the entire machinery for the generation of peptides
00:34:21;03 is located in the cytoplasm, whereas the molecule charged with antigen presentation
00:34:26;14 lives in extracellular space.
00:34:28;18 So somehow we must deliver peptides to extracellular space.
00:34:32;18 And this is the function of a dedicated transporter referred to as the transporter associated
00:34:37;23 with antigen presentation, or the TAP protein, indicated by this array of helical segments here.
00:34:47;03 Once peptides are translocated into the endoplasmic reticulum, they become part of
00:34:52;01 a nascent class-I MHC product, which itself requires the action of a panoply of chaperones to ensure its
00:35:00;04 proper folding.
00:35:01;04 But when all is said and done, we make this peptide-MHC complex, which is then free
00:35:05;09 to travel to the cell surface.
00:35:07;01 And as I've suggested in the preceding slide, viruses are masters of deception.
00:35:12;07 They've evolved numerous countermeasures with which to frustrate this process of antigen presentation.
00:35:18;09 And here's just an example taken from herpes viruses, one class of pathogens that
00:35:23;09 once you acquire them stay with you for the rest of your life.
00:35:26;26 We have proteins that in... such as pp65 that involve... that interfere with ubiquitylation
00:35:34;25 of possible targets.
00:35:37;24 The virus that is the causative agent of mononucleosis, Epstein-Barr virus, produces a protein
00:35:43;22 that renders viral products insensitive to proteolytic digestion by the proteasome.
00:35:49;03 We have other herpes virus-encoded proteins that impede peptide translocation into
00:35:53;08 the endoplasmic reticulum, detain class-I molecules at the site of synthesis,
00:35:58;23 or even reverse the process of membrane insertion and target those very same MHC products
00:36:03;19 for proteasomal degradation.
00:36:06;01 The process is more complex than this.
00:36:09;04 We have meanwhile figured out some of the details.
00:36:11;14 This is the mating dance between the viral protein US2 and the class-I molecule it destroys.
00:36:17;26 And in a process referred to as retrotranslocation, a newly assembled class-I heavy chain is
00:36:23;16 sent back to the cytoplasm for proteasomal degradation.
00:36:27;01 This is just one example of the many tricks viruses can use to frustrate adaptive immunity.
00:36:33;02 And such interference may apply to other surface proteins, cytokines released from the cell,
00:36:39;10 aspects of innate immunity.
00:36:40;11 I need to emphasize the fact that the constant interplay between the immune system,
00:36:45;27 which exerts a selective pressure, and pathogens, which have the capacity to rapidly evolve,
00:36:51;22 results in this per... perpetual chess game between host and pathogen.
00:36:57;13 Much of this work enables cell biological explorations that would be difficult to achieve otherwise.
00:37:03;17 And to put some molecular detail on this particular cartoon, this would be our current understanding
00:37:09;04 of how this complicated machine operates.
00:37:11;25 We have this centrally positioned class-I MHC product and a host of other cofactors
00:37:17;12 that together ensure that this class-I protein in a virus-infected cell can be extracted
00:37:22;24 from the endoplasmic reticulum and ultimately targeted for proteasomal degradation.
00:37:27;27 So, after this whirlwind tour of the immune system, let me return to where we started.
00:37:33;19 We have a multi-layered immune defense system, of which the mechanical and innate immune defenses
00:37:39;13 are probably the most important on a daily basis.
00:37:42;20 But once these systems fail, adaptive immunity kicks in.
00:37:46;26 And the remarkable precision with... with which the adaptive immune system can recognize antigens
00:37:51;23 has allowed the explorations which I've tried to summarize in the preceding
00:37:56;17 thirty minutes or so.
00:37:58;02 Key features: ability to distinguish between structures that differ by very little -- as few as an atom, perhaps;
00:38:06;06 the ability to respond rapidly;
00:38:09;25 and the ability to adjust the specificity of the ensuing response to whatever the needs of the day may be.
00:38:17;14 In the second part of my talk, I will highlight one specific element of this adaptive immune system.
00:38:23;22 And we'll see how this can be leveraged into tools that might be useful,
00:38:28;24 both for basic cell biology as well as for biomedicine.

Part 2: Unusual Antibody Fragments: The Camelid Nanobodies

00:00:07;24 My name is Hidde Ploegh.
00:00:08;24 I'm a Senior Investigator at Boston Children's Hospital in the program of Cellular and Molecular Medicine.
00:00:15;19 And in this part of my presentation, I'll discuss some more recent work that describes
00:00:22;05 efforts to harness unusual features of antibodies produced by camelids.
00:00:30;01 Why would this be interesting?
00:00:34;07 These are the animals that we work with.
00:00:36;06 These are two juvenile alpacas, Bryson and Sanchez.
00:00:40;11 And the reason why these animals attracted my attention was spawned by a lecture that
00:00:45;07 I attended, in which I was misassigned to the wrong section.
00:00:48;13 And it's the first time I heard about the unusual properties of antibodies made by camelid species,
00:00:53;22 and that includes alpacas, llamas, guanacos, vicunas, camels, dromedaries --
00:01:00;05 all camelids shared this unique feature.
00:01:04;01 What I'm talking about is the fact that, unlike the typical mammalian species, which uniquely
00:01:09;05 produce antibodies composed of heavy chains and light chains, camelids also produce,
00:01:15;01 in addition to the traditional forms, heavy-chain only antibodies.
00:01:18;16 And this is of relevance because for many applications the size of a full-sized antibody
00:01:25;20 is incompatible with some of the applications we seek for them.
00:01:30;02 And it is for this reason that we try to shrink intact antibody molecules to the minimal unit
00:01:35;17 that can recognize antigen.
00:01:37;24 As I discussed in the preceding presentation, this can be done by proteolytic digestion.
00:01:44;11 Using papain, we can produce so-called Fab fragments composed of the variable regions
00:01:49;05 of the heavy chain and the light chain and one of their constant domains, respectively.
00:01:54;05 These retain antigen binding properties, but it requires considerable effort to
00:01:59;03 produce them in a proper way.
00:02:01;23 Your friend, the molecular biologist, can help you shrink these Fab fragments even further,
00:02:06;10 to create so-called single-chain Fv fragments, in which the variable regions of
00:02:11;11 the heavy chain and light chain are connected by a linker.
00:02:14;28 The problem that often arises with expression of these single-chain Fv fragments is that
00:02:20;12 they have not been through the secretory pathway of a eukaryotic cell, and consequently
00:02:25;07 they tend to be aggregation-prone and may require considerable optimization before you end up
00:02:30;13 with a stable product.
00:02:33;02 The heavy-chain only antibodies made by the camelids have no such disadvantages because
00:02:38;05 you can shrink the recognition module to just the variable region of the heavy chain.
00:02:42;20 And unlike what applies to the traditional antibodies, all features of structure required
00:02:48;14 for specific recognition of antigen are embedded within the variable regions of the heavy chain.
00:02:54;09 These fragments are called VHH fragments and are commonly referred to as nanobodies as well.
00:03:01;06 Some of the properties that make these nanobodies unusually attractive include the fact that
00:03:07;06 they can be produced in bacteria in high yield.
00:03:10;12 Many of them require neither glycosylation nor disulfide bonds for stability.
00:03:14;13 And their small size enables applications for which even single-chain Fvs would be
00:03:19;27 hard up to the task.
00:03:21;15 So, how do we make these single-chain variable region fragments?
00:03:26;15 We practice what is called multiplexed immunization, a fancy word for sticking up to two dozen
00:03:31;12 different antigens into an animal.
00:03:33;27 And after repeated boosting of the animal to achieve high affinity antibodies, we isolate
00:03:39;25 lymphocytes from peripheral blood.
00:03:42;04 And using primers designed to uniquely amplify the variable regions of
00:03:46;24 these heavy-chain only antibodies, we can create an immortalized collection of these VHH sequences.
00:03:55;03 How does one identify VHH sequences of interest?
00:03:59;04 This is done by cloning them in a phage display vector, where one can select phage that express,
00:04:07;03 as a fusion of the P3 protein, of the phage coat, a nanobody or VHH fragment.
00:04:13;06 And by performing what is called a panning reaction, by immobilizing antigen on a solid support,
00:04:18;16 we can retrieve those phage particles that express the nanobody of interest,
00:04:24;06 specific for the antigen with which we coated the plate.
00:04:28;05 So, even though we may start with animals immunized with two dozen different antigens,
00:04:32;08 by screening them individually on these antigen-coated plates we end up with nanobodies of
00:04:38;12 defined and unique specificity, which we may then produce in bacteria in high yield.
00:04:44;11 This is the conventional means of screening these so-called nanobodies.
00:04:48;28 But we've considered also the fact that these nanobodies require neither glycosylation
00:04:54;17 nor disulfide bonds, which enables their expression in the reducing environment of the cytosol.
00:05:00;06 And this has allowed us to perform functional screens to select for nanobodies that
00:05:05;11 perturb reactions inside the cytoplasm of, in my example, a mammalian cell.
00:05:10;24 So, let me give you an example.
00:05:12;10 Taking advantage of the fact that these nanobodies often require neither glycosylation nor disulfide bonds,
00:05:18;18 we devised a screening procedure that relies on the cytoplasmic expression of these nanobodies
00:05:25;09 to impede virus infection.
00:05:28;20 In this particular case, we took an animal and immunized it with disrupted influenza virus
00:05:34;03 and vesicular stomatitis virus, abbreviated as IAV and VSV, respectively, in the ensuing slides.
00:05:42;11 And upon immunization, and having extracted the RNA, instead of cloning the PCR products
00:05:47;27 into a phage display vector, as shown in the preceding slide, we clone these
00:05:52;28 amplified products in a lentiviral vector such that expression of the nanobody is under the control
00:05:58;06 of a doxycycline-inducible promoter.
00:06:01;14 Cells transduced with a lentiviral vector will produce the nanobody only when
00:06:06;00 properly exposed to doxycycline.
00:06:08;02 So, we transduce a collection of mammalian cells with these lentiviral vectors such that,
00:06:14;09 on average, each cell will receive a single antiviral construct.
00:06:18;25 Upon induction with doxycycline, that cell will produce the corresponding nanobody
00:06:23;13 in its cytoplasm, and we can then test this nanobody for its functional properties.
00:06:28;18 By exposing the transducants, as an entire collection, to a lethal challenge
00:06:33;19 -- in this case, a lytic dose of either influenza virus or vesicular stomatitis virus -- we can select
00:06:39;15 for survivors, the underlying logic being that any nanobody that impedes an aspect of
00:06:44;20 the virus replication cycle will confer a growth advantage to that cell.
00:06:49;24 We can then take the survivors and rescue the VHH or nanobody sequences, and then
00:06:55;06 validate them by independent means.
00:06:57;22 So, this is an example of such a screen, where we tested individual nanobody-producing clones
00:07:06;05 in the absence and presence of doxycycline induction.
00:07:09;11 You'll see two bars for each cell line analyzed.
00:07:13;12 The open bars indicate the situation in the absence of doxycycline induction, and we
00:07:19;11 monitor the infected state by independent means.
00:07:21;22 And as the open bars show, across the board in the absence of doxycycline induction,
00:07:26;27 each of these transduction... transductants is fully infectable by influenza virus.
00:07:32;21 But as the solid bars show, there are numerous clones, selected in this case as
00:07:37;23 protecting against influenza infect... infection, that drastically reduced the ability of the virus
00:07:42;27 to replicate.
00:07:43;28 When we test those very same clones in response to an infection with vesicular stomatitis virus,
00:07:49;00 clones that are fully protected against influenza virus are not responsive to infection
00:07:56;02 with VSV at all.
00:07:57;11 It doesn't matter whether we induce expression of the nanobody;
00:08:00;14 infection with the irrelevant virus proceeds normally.
00:08:03;26 We can conduct the converse experiment by selecting for clones that impede VSV replication,
00:08:09;05 and these in turn leave replication of influenza virus un... untouched.
00:08:14;04 Now, one of the appealing features of nanobodies, and this is one of the reasons why structural biologists
00:08:20;05 have warmed up to them, is the fact that they often serve as crystallization chaperones.
00:08:25;24 That means that for proteins that are often challenging to crystallize, inclusion of
00:08:30;20 a nanobody of the appropriate specificity will facilitate crystallization.
00:08:35;01 This of course has the tremendous advantage of identifying at atomic resolution
00:08:40;03 exactly the epitope to which the nanobody in question binds.
00:08:44;01 And this was done by Florian Schmidt and Leo Hanke, two students and postdocs in my lab.
00:08:50;27 They together solved the crystal structure of, in this case, the target of the anti-influenza
00:08:58;14 nanobodies which impede replication in complex with a representative nanobody.
00:09:06;03 In red and orange are indicated those sequences in the flu nucleoprotein that are responsible
00:09:13;08 for interaction with RNA -- because influenza virus is a negative-stranded RNA virus --
00:09:19;00 and in gold the nuclear localization signal.
00:09:22;04 And what this experiment reveals is that the nanobody binds to an area of the NP protein
00:09:27;04 not previously implicated as functionally important.
00:09:30;17 So in that sense, having this nanobody provides us with a novel angle of attack
00:09:36;00 on the virus life cycle.
00:09:37;01 Now, if one asks, by what molecular mechanism could this nanobody conceivably impede virus replication?,
00:09:44;09 we made use of a structure published for the viral ribonucleoprotein particle.
00:09:50;09 This is a viral RNA in complex with nucleoprotein.
00:09:54;13 The analogy might be of a histone binding to DNA.
00:09:57;26 And what we found is that the nanobody, when bound to the nucleoprotein, occludes the
00:10:03;04 nuclear localization signal on an adjacent NP subunit, and thus impedes import of the viral
00:10:10;27 ribonucleoprotein particle into the nucleus, which is a step essential for flu replication.
00:10:17;13 Similar experiments were done for the VSV-inhibitory nanobodies.
00:10:21;14 And in this particular case we again found that the nanobody binds to the N protein,
00:10:27;13 which is the conceptual equivalent of the NP protein in influenza virus.
00:10:32;12 The N protein complexes RNA, barely visible here as this red material found in between
00:10:38;06 the two lobes of the N protein.
00:10:40;19 The N protein forms a decameric ring.
00:10:44;04 And in the crystal structure solved for the complex between the N protein and the nanobody,
00:10:50;04 we find the presence of two such decameric rings, each subunit of which is crowned by
00:10:55;02 an individual nanobody as part of the crystal structure.
00:10:58;16 So, we have 40 subunits to the asymmetric unit.
00:11:02;02 And this is one of the largest protein structures to which my lab has contributed.
00:11:05;08 This was done together with Tom Schwartz at MIT.
00:11:10;02 If we now ask, would this particular structure help us understand exactly how the nanobody
00:11:16;03 impedes virus replication?, the answer is yes.
00:11:19;08 It turns out that the nanobody binds to a segment of the N protein to which also binds
00:11:24;08 the polymerase cofactor P.
00:11:27;09 In order for the VSV genome to be replicated,
00:11:29;28 the polymerase must interact with cofactor P to enable that reaction to proceed.
00:11:35;14 And as we observed, the nanobody competes for binding with this P protein, thus adequately
00:11:40;18 explaining why it inhibits virus replication.
00:11:44;18 And what is not shown in this slide is the fact that if we now grow virus in the presence
00:11:49;00 of this particular nanobody, we can readily select for escape variants that no longer
00:11:54;04 are susceptible to inhibition by this nanobody.
00:11:57;14 And as you might expect, the mutations that account for these escape variants map exactly
00:12:02;14 to the site to which the nanobody binds without affecting binding of cofactor P.
00:12:08;08 So we have here a complete cycle, in which we can immunize an animal with some antigenic preparation,
00:12:15;06 conduct a phenotypic screen, identify the epitope recognized at atomic resolution,
00:12:20;23 and then use an in vivo biological experiment to validate these conclusions by selecting
00:12:25;17 for escape variants.
00:12:28;17 A second example that exploits the ability of nanobodies to function in the cytoplasm
00:12:34;20 of a mammalian cell concerns nanobodies that we've made against the component
00:12:39;26 of a highly complex structure we refer to as the inflammasome.
00:12:43;14 And without going into unnecessary detail, inflammasomes are supramolecular assemblies
00:12:49;06 required for a particular aspect of innate immunity.
00:12:52;25 When an inflammasomes senses a microbial component, it oligomerizes, and it makes use of
00:13:00;09 a series of adapters to then recruit massive numbers of an enzyme called pro-caspase-1.
00:13:06;15 Once this inflammasome is properly assembled, it leads to activation of pro-caspase-1,
00:13:13;06 and this in turn allows conversion of a cytokine called pro-IL-1 into its active counterpart,
00:13:19;26 IL-1.
00:13:20;26 This is a key aspect of inflammation.
00:13:23;22 It turns out that inflammasomes occur in many molecular manifestations, two of which
00:13:29;04 are shown here.
00:13:30;04 On the left, the NLRP3 inflammasome, composed of leucine-rich repeats
00:13:35;06 -- those are these repetitive structures shown on top;
00:13:38;28 a pyrin domain via which it recruits the adapter ASC;
00:13:42;18 and the ASC protein itself has yet another protein-protein interaction domain called
00:13:46;12 a caspase activation and recruitment domain via which, as the name suggests,
00:13:51;04 caspase is recruited.
00:13:53;21 These inflammasomes occur in many different flavors.
00:13:56;08 A second one, not the only one that exists, is the NAI...
00:14:02;05 NAIP or NLRC4 inflammasome, where the involvement of the ASC adapter was considered under debate.
00:14:12;22 So, we raised nanobodies against the ASC protein, found that these antibodies could in fact
00:14:20;17 impede proper release of interleukin-1, suggesting that indeed we had an inhibitory antibody,
00:14:28;06 and again, by exploiting the fact that nanobodies serve as crystallization chaperones,
00:14:34;04 we were lucky enough to be able to determine exactly how this particularly nanobody binds to
00:14:38;23 its counterstructure, the CARD domain of the ASC protein.
00:14:43;04 What is noteworthy in this particular case is that the nanobody binds to its target
00:14:48;08 not via the complementarity determining region 3, which most often makes the key contribution
00:14:53;07 to antigen binding, but with a very important contribution of CDR2 to mediate specific interaction
00:14:59;23 with the CARD domain.
00:15:01;06 And this was work done in collaboration with Hao Wu at Boston Children's Hospital.
00:15:06;08 Now, having a nanobody that binds to this ASC protein enables us to do a variety of
00:15:13;22 different experiments.
00:15:15;05 And one such experiment involves the construction of a sensor that might report on the assembly
00:15:21;12 of inflammasome monomers and ASC monomers into the polymeric structures that are a prerequisite
00:15:27;18 for activation.
00:15:30;06 By fusing our nanobody specific for ASC to green fluorescent protein, for the first time
00:15:37;22 we can look at the distribution of this ASC protein at natural abundance.
00:15:41;22 One of the complications of inflammasome study is that overexpression of individual components
00:15:47;25 can lead to art... artifactual and premature activation.
00:15:51;13 And what we would like to know is what the behavior of these components looks like
00:15:55;26 under conditions of natural abundance.
00:15:58;03 So in this particular case, we took a monocytic human cell line capable of sustaining NLRP3 activation,
00:16:06;25 we introduced the reporter composed of the ASC-specific nanobody fused to GFP,
00:16:12;15 and in the absence of overt stimulation we looked for the distribution of ASC protein
00:16:17;20 by virtue of the fact that it's targeted by our nanobody-GFP fusion construct.
00:16:23;06 And what you see in the absence of any overt activation is this diffuse distribution
00:16:28;17 of the ASC protein.
00:16:30;16 If we now expose those very same cells to inflammasome triggers, we can begin to see
00:16:35;09 the initiation of polymerization.
00:16:36;20 We see the formation of these filaments.
00:16:39;28 The assembly of the inflammasome does not proceed to completion, which explains
00:16:44;12 the inhibitory effect of this particular nanobody, but it gives you an example of how we can
00:16:49;14 use nanobodies to report, in live cells, on phenomena that would otherwise be very difficult
00:16:56;10 to study.
00:16:58;21 Another key application which we've been interested in for a long time is to use
00:17:03;02 single-domain antibody fragments or nanobodies for imaging purposes.
00:17:07;18 If you want to develop an antibody-type reagent for imaging, it needs to meet
00:17:11;23 a certain set of criteria.
00:17:14;08 It needs to be labeled so we can detect the labeled product.
00:17:17;28 It needs to be cleared from the circulation rapidly to achieve proper signal-to-noise ratio.
00:17:23;23 And it needs to achieve adequate tissue penetration.
00:17:26;16 And this is where its small size becomes a key distinguishing feature in comparison
00:17:31;13 with full-sized antibodies.
00:17:34;00 Full-sized antibodies are about 150 kiloDaltons in molecular mass.
00:17:38;24 The nanobody fragments with which we work are approximately one-tenth the size.
00:17:42;16 And this accounts for their superior tissue penetration, their accelerated clearance
00:17:47;07 from the circulation, and thus we thought these would be ideal agents for imaging purposes.
00:17:52;04 Now, in order to label an antibody fragment, or any protein for that matter, you want to
00:17:58;20 avoid collateral damage.
00:18:01;20 Most of the chemical labeling procedures that are commonly used entail the risk of modification
00:18:07;17 of residues important for function of that protein.
00:18:10;00 And so, in the design of our imaging agents, we wish to avoid the introduction of
00:18:15;17 any chemical damage to the protein of interest.
00:18:18;22 So, to achieve that, we make use of what's called a sortase reaction.
00:18:24;10 The sortase enzyme is a transacylase that recognizes a short peptide sequence
00:18:31;07 at or close to the C-terminus of any target protein of interest.
00:18:35;28 It's essentially a thiol protease that cleaves between the threonine and glycine,
00:18:40;20 indicated in single-letter code in this slide, to result in a covalent thioacyl enzyme intermediate,
00:18:47;11 indicated in the bottom left, here.
00:18:48;27 Now, we're providing, ex vivo, a short synthetic peptide, minimally composed of a single glycine
00:18:57;12 -- ideally, two, three or more -- to which one may affix pretty much any payload of interest,
00:19:03;18 denoted here as probe.
00:19:05;14 These synthetic peptides serve as the incoming nucleophiles that attack this thioacyl enzyme
00:19:10;15 intermediate.
00:19:11;23 And it results in the formation of a covalent bond between the target protein that you've
00:19:16;17 modified with this LPXT(G)n sequence in perfectly good peptide linkage with the probe,
00:19:24;13 attached via this oligoglycine stretch.
00:19:26;26 So, this is a method that lends itself to modification of pretty much any protein of interest,
00:19:32;00 if properly optimized.
00:19:34;18 Now, the types of payloads that one can attach to proteins are limited only by what
00:19:40;24 your friend, the chemical... the chemists can produce.
00:19:44;05 In this example, I've indicated a few of the things we've successfully been able to affix
00:19:49;00 using this sortase reaction.
00:19:51;05 It includes site-specific biotinylation, the site-specific and quantitative attachment
00:19:56;19 of fluorophores, execution of protein-protein fusions, attachment of DNA molecules, and,
00:20:03;00 as will be relevant for what I have to say in the following, the attachment of radioisotopes.
00:20:09;26 The important thing to remember here is that the reaction proceeds site-specifically and
00:20:15;03 nearly stoichiometrically, which makes it very attractive for protein engineering purposes.
00:20:20;23 So, in order for us to use nanobodies against surface proteins of interest for imaging purposes,
00:20:29;24 we need to install a label that allows us to detect them.
00:20:33;21 And without... without going into the physics, we've been interested in positron emission tomography.
00:20:39;08 This is a method that can be used in the clinic to diagnose metabolically active from inactive cells.
00:20:45;01 It's a non-invasive method.
00:20:47;02 But it... it uses radioactive isotopes that are sometimes challenging to handle.
00:20:52;05 Without going into the chemical details, we've developed procedures to affix these isotopes
00:20:56;25 suitable for PET imaging to our nanobodies using the sortase reaction.
00:21:01;20 So, on the left you see some nasty chemistry that results in the formation of an F-18 labelled compound.
00:21:09;26 Fluorine 18 is the isotope suitable for PET imaging.
00:21:13;12 On the right, you see the workstream for our sortase-catalyzed reaction to install
00:21:17;27 a chemical entity that now allows... allows us to combine the two under strictly native conditions,
00:21:24;00 without inflicting any collateral damage onto the protein of interest.
00:21:28;00 So, this is how we can make nanobodies ready for imaging by positron emission tomography.
00:21:33;27 And what is shown in this little inset, here, is a mouse that we imaged with an antibody
00:21:38;00 against class-II MHC products.
00:21:41;02 These are molecules found on antigen-presenting cells.
00:21:43;28 In this case, we deal with a tumor bearing mouse where the presence of antigen-presenting cells
00:21:49;20 in the tumor can be visualized by this non-invasive imaging method.
00:21:56;18 We've also been interested, in particular, in imaging immune responses non... non-invasively.
00:22:03;05 I pointed out the fact that T lymphocytes recognize antigen by virtue of this antigen-specific
00:22:09;18 receptor that recognizes MHC products complexed with the appropriate peptide, but the reactivity
00:22:16;06 of these T cells can be mon... monit... can be modulated by the expression of
00:22:21;06 so-called costimulatory or checkpoint molecules.
00:22:23;21 And that can be either stimulatory or inhibitory.
00:22:26;10 And in order for us to image immune responses non-invasively, to focus on
00:22:31;21 these costimulatory or checkpoint molecules became a point of interest.
00:22:37;11 This is all the more important when we now know that these checkpoint molecules are important
00:22:42;19 targets for immunotherapy of cancer.
00:22:46;12 This slide is taken from the work of Jim Allison, who showed that by blocking one of these checkpoint
00:22:51;20 molecules one could achieve a remarkable improvement in an antitumor response.
00:22:58;01 In this case, you see what happens to tumor growth, measured as tumor diameter.
00:23:03;28 In the absence of any treatment, if you give an antibody that fails to impede this regulatory interaction,
00:23:12;11 no effect on tumor growth.
00:23:14;18 But if you block a checkpoint molecule that confers inhibition on T cells of
00:23:19;26 appropriate specificity, merely blocking that inhibitory molecule gives you a significant improvement
00:23:25;26 in the antitumor response.
00:23:27;20 Now, this has been wildly successful in the clinic, cures having been achieved, but only
00:23:33;19 a fraction of patients respond to such treatment and it will be very important to know
00:23:38;06 why certain people do respond to therapy and others do not.
00:23:42;12 And we think that the ability to image immune responses non-invasively may help us
00:23:47;01 provide some answers to that question.
00:23:50;03 In addition to the CTLA-4 protein, which was the focus of attention of work in the Allison lab,
00:23:55;10 much work has since focused on other numbers of the checkpoint family,
00:24:01;12 notably the PD-L1 protein.
00:24:03;02 So, let me take you through this example.
00:24:05;17 We have here a T lymphocyte that engages an antigen-presenting cell via its T cell receptor
00:24:11;08 for an antigen.
00:24:12;26 But if this antigen-presenting cell also expresses PD-L1, it engages a counterstructure
00:24:19;08 on the T cell which now confers an inhibitory signal to that T cell rendering it incapable of
00:24:25;05 exerting its usual functions, be that cytokine production or a lytic function.
00:24:30;18 If one has antibodies that block either PD-1 or PD-L1, such inhibition may be relieved.
00:24:37;11 And just like the anti-CLTA-4 example of which I showed you the graph, blocking PD-1 or PD-L1
00:24:43;17 has been shown to have a significant antitumor effect.
00:24:48;00 So, we developed nanobodies against PD-L1.
00:24:53;05 And we're able to show that the B16 melanoma, which is a frequently used mouse model of melanoma,
00:24:59;09 expresses the PD-L1 protein.
00:25:02;13 What you see here are mice injected with a radiolabeled fragment of a nanobody that recognizes
00:25:07;28 PD-L1.
00:25:09;22 These mice on their left shoulder carry a tumor.
00:25:13;15 And the presence of the PD-L1 protein on the tumor may be clearly revealed by this non-invasive
00:25:19;22 imaging method.
00:25:21;10 We always use genetic controls to infer specificity of the labeling reactions that we see.
00:25:27;17 On the right, you see a PD-L1 knockout mouse inoculated with the same tumor.
00:25:32;20 And although somewhat more faint, we can still detect the PD-L1 signal in this setting as well.
00:25:38;17 The conclusion from this experiment is that PD-L1-positivity in the tumor microenvironment
00:25:44;05 is the net result of expression of that protein on invading cells of hematopoietic origin,
00:25:50;28 as well as by expression of PD-L1 by the tumor itself.
00:25:55;04 So, while it is possible to use a radioactive isotope of fluorine, F-18, for labeling purposes,
00:26:02;22 its short half-life, 110 minutes, poses limitations.
00:26:06;28 For that reason, fortunately we have access to other isotopes that are PET-compatible,
00:26:11;07 such as zirconium-89.
00:26:13;17 And we have the chemical means to install those site-specifically as well.
00:26:18;05 In panel A, we indicate the chemical structure of the chelator required to specifically bind
00:26:25;01 zirconium-89.
00:26:27;00 And in much the same way as we use sortase to install F-18, we have chemistry that
00:26:31;21 enables us to install these zirconium-labeled compounds.
00:26:35;22 The bottom half of this slide illustrates the basic enzymatic approach we follow to
00:26:40;20 attach this zirconium.
00:26:42;19 Again, we use a sortase reaction to install the structure indicated in panel A.
00:26:47;27 And we have the added capability of modifying these structures even further through installation
00:26:54;02 of polyethylene glycol, for example, to affect the hydrophobic-hydrophilic balance of
00:26:59;16 the final product.
00:27:00;18 So, here you see an experiment in which we took a nanobody that recognizes CD8.
00:27:06;26 This is the glycoprotein marker diagnostic of cytotoxic T lymphocytes.
00:27:10;18 And on the far left you see the initial result with a nanobody simply labeled with zirconium-89.
00:27:18;10 The two massive structures seen on the bottom are the kidneys and the bladder.
00:27:22;26 And this particular nanobody, when labeled, tends to concentrate in the organs of elimination,
00:27:28;07 with only very faint labeling visible in the thymus, the lymph nodes
00:27:31;27 -- exactly where you would expect it.
00:27:34;20 But by now adding PEG spaces of 5, 10, and 20 kiloDaltons, thus improving hydrophilicity
00:27:43;03 of the labeling agent, we see a massive reduction in this uptake in the organs of elimination.
00:27:49;17 And we begin to see, with great clarity, the lymphoid structures these nanobodies
00:27:53;19 were designed to detect in the first place.
00:27:55;28 Again, we use a genetic control.
00:27:59;00 In the far right, you see a mouse that lacks any and all lymphocytes, owing to the fact
00:28:04;13 that it's mutated in the very locus responsible for VDJ rearrangement.
00:28:10;18 And the lymphoid structures so clearly visible in this panel are altogether absent in the
00:28:16;20 RAG knockout, which lacks lymphocytes altogether.
00:28:19;28 So, with this tool in hand, we can finally begin to image cytotoxic T lymphocytes
00:28:25;04 and ask the question, can we detect the influx of killer T cells into the tumor microenvironment?
00:28:32;11 What this slide shows is a mouse that's been inoculated with a pancreatic tumor cell line.
00:28:38;10 And on its left flank, you can clearly see the tumor by virtue of the fact that CD8 lymphocytes
00:28:45;15 infiltrate it, showing that we can indeed use this kind of non-invasive imag... imaging
00:28:51;08 to track the behavior of T lymphocytes in living mice.
00:28:54;18 And what's really important about this type of experiment is the fact that we no longer
00:28:58;16 need to kill the animal to examine what goes on inside it.
00:29:02;23 Typically, immunologists, in order for them to access lymphoid organs or to access lymphocytes
00:29:09;08 that reside in the tumor, have to kill the mouse, excise the organs of interest,
00:29:14;17 and then enumerate and characterize lymphocytes that are present.
00:29:18;10 With this non-invasive imaging method, albeit at modest resolution, we can now track
00:29:23;13 an animal over time and watch the immune response unfold.
00:29:27;13 So, this is our famous spinning mouse.
00:29:30;18 And once again, you can clearly see the tumor on the right flank.
00:29:35;11 In fact, you will also see that these dots represent the lymph nodes.
00:29:40;24 And lymph nodes that are in close proximity to the tumor are more intensely labeled
00:29:46;00 than lymph nodes... lymph nodes on the other side, consistent with trafficking of lymphocytes
00:29:50;18 between these different lymphoid organs.
00:29:52;26 So, how can we use this information to address one of the questions I posed earlier?
00:29:59;05 Why is it that certain tumors respond to checkpoint-blocking antibodies,
00:30:03;18 such as anti-CTLA-4 or anti-PD-1/PL-L1, and others do not?
00:30:09;03 Can we use this particular imaging method to shed some light on what happens in
00:30:13;21 that particular setting?
00:30:15;14 So, this is a complicated slide, but the concept is very simple.
00:30:20;27 We use our B16 melanoma model in a setting where some of the animals will proceed
00:30:27;03 to full-blown tumor growth and others will respond to control of the tumor in response to
00:30:33;22 checkpoint-blocking antibodies.
00:30:36;24 So in this case, we use anti- CTLA-4 antibodies, full-sized antibodies, as a therapeutic agent.
00:30:43;01 And we simply image in the tumor microenvironment what happens to the CD8 T lymphocytes
00:30:48;02 that one finds on location.
00:30:49;22 And what we find is a very strong correlation between the distribution of these CD8 T cells
00:30:56;03 and the ability of an animal to control the tumor.
00:30:59;06 If we see monofocal distribution of T lymphocytes, this is paralleled by a reduction in tumor size
00:31:05;20 and improved survival of the animals.
00:31:08;24 In animals that fail to respond to checkpoint blockade, and where the tumors grow,
00:31:13;10 we see this more heterodisperse distribution of CD8 T cells, and this becomes a predictor,
00:31:18;22 if you will, of the outcome of the final experiment.
00:31:23;19 This was validated, in collaboration with the lab of Bob Weinberg, in a breast cancer model
00:31:27;11 where two variants exist.
00:31:30;10 One is an epithelial variant that is immunogenic and is readily controlled in response to checkpoint blockade,
00:31:36;10 and there's a mesenchymal variant which is less differentiated, highly aggressive,
00:31:41;01 and poorly responsive to checkpoint blockade.
00:31:44;04 If we now image animals of either type... you see an example of the mesenchymal variant,
00:31:49;09 poorly controlled by checkpoint blockade... tumors continue to grow, and the animals
00:31:54;06 do not survive.
00:31:55;10 On the bottom, you see the epithelial variant, a monofocal distribution of CD8 T cells,
00:32:02;04 excellent response to checkpoint-blocking antibodies, and thus we think that this method for the
00:32:06;20 first time begins to shed some light on what happens in the tumor microenvironment.
00:32:10;24 We have a long way to go and there's a long list of surface molecules that we'd like to image.
00:32:16;18 There's a lot that we need to know about various aspects of the immune response.
00:32:20;15 But it will be important to measure these things in the context of a living experimental animal
00:32:25;19 and ultimately, perhaps, in the clinic.
00:32:27;25 So, the preceding segment has shown our efforts in imaging a certain aspect of adaptive immunity,
00:32:34;17 notably CD8 T cells.
00:32:37;11 But as I pointed out, host defense is an integrated system that comprises not only adaptive immunity
00:32:44;09 but innate immunity as well.
00:32:45;25 And future efforts are focused on trying to image these innate immune responses non-invasively
00:32:51;18 as well.
00:32:52;18 I think inflammation has a key aspect that would benefit from this approach.
00:32:57;12 The nanobodies that we've used have a few signature advantages:
00:33:00;28 their small size makes imaging possible;
00:33:03;20 ease of manufacture in E. coli is a plus;
00:33:06;13 the fact that they require neither glycosylation nor disulfide bonds allows us to express them
00:33:12;01 in the cytoplasm of mammalian cells and use them to conduct phenotypic screens, as I've demonstrated;
00:33:19;00 finally, their small size makes it possible to exploit the sortase reaction to
00:33:24;03 install pretty much any... any entity of choice without inflicting damage on the nanobody itself,
00:33:29;03 and that is the method we have applied to render nanobodies suitable for imaging by
00:33:35;06 positron emission tomography.
00:33:37;25 Let me finish by acknowledging the individuals who have done this work.
00:33:41;23 The virus work that I summarized was done by Florian Schmidt, a postdoc,
00:33:47;09 and Leo Hanke, a graduate student.
00:33:50;17 The nanobody platform was established by Jessica Ingram, currently on the faculty of Dana-Farber.
00:33:56;11 The radiochemistry was done by Mohammed Rashidian.
00:33:59;10 And the tumor immunology experiments were conducted with very the significant input
00:34:04;06 and participation of Mike Dougan.
00:34:07;05 I've mentioned the fact that we collaborate with Bob Weinberg on our breast cancer models,
00:34:11;13 Hao Wu on the inflammasome,
00:34:14;24 Steve Almo as a source of proteins with we... with which
00:34:17;03 we can immunize alpacas and obtain the antibodies with which we work.
00:34:21;07 Thank you very much.

Videos in this Talk

Part 1: Immunology: The Basics of Antibody Diversity
Audience:
- Student
- Researcher
- Educators
- Educators of H. School / Intro Undergrad
- Educators of Adv. Undergrad / Grad
Part 2: Unusual Antibody Fragments: The Camelid Nanobodies
Audience:
- Researcher
- Educators of Adv. Undergrad / Grad

Speaker: Hidde Ploegh

Total Duration: 1:13:32

Recorded: July 2017

All Talks in Immunology

Talk Overview

How does our immune system protect us against all of the infectious agents and foreign substances we encounter? Much of the answer lies in antibody diversity. In his first talk, Dr. Hidde Ploegh explains how B cells shuffle their genetic material such that regions of the immunoglobulin protein are rearranged. This generates the antibody diversity needed to recognize an almost infinite number of antigens. Interactions of B cells with T helper cells results in the formation of structurally distinct classes of immunoglobulins, further increasing antibody diversity. T killer cells are primed to attack infectious agents when immunoglobulins on their surface recognize antigens presented by the major histocompatibility complex (MHC). Ploegh explains that by subverting the MHC pathway, viruses and cancer cells can evade the immune system.

In part two, Ploegh describes how his lab takes advantage of the unique properties of antibodies from the Camelidae family (alpacas, llamas, camels, etc). In addition to traditional antibodies, these animals naturally make small, heavy-chain only antibodies (nanobodies). These molecules can be isolated, amplified in bacteria, and engineered for new applications. As well as using nanobodies to target viruses and inflammasomes, Ploegh explains how his lab uses labelled nanobodies for non-invasive, live imaging of cancer tumors in mice. These technologies have exciting implications in basic and biomedical studies.

Speaker Bio

Hidde Ploegh

Dr. Hidde Ploegh is an immunologist at Boston Children’s Hospital. His love for immunology began when he was an undergraduate at the University of Groningen in the Netherlands. As a student, he wrote a letter to renowned immunologist Jon van Rood but never heard back. However, as an undergraduate researcher, he had an opportunity to… Continue Reading