免疫学（英文版）：Chapter 05

chapter 5 DNA. While we think of genomic DNA as a stable genetic blueprint, the lymphocyte cell lineage does not retain an in- tact copy of this blueprint. Genomic rearrangement is an es- sential feature of lymphocyte differentiation, and no other vertebrate cell type has been shown to undergo this process. This chapter first describes the detailed organization of the immunoglobulin genes, the process of Ig-gene rearrange- ment, and the mechanisms by which the dynamic im- munoglobulin genetic system generates more than 10 8 different antigenic specificities. Then it describes the mecha- nism of class switching, the role of differential RNA process- ing in the expression of immunoglobulin genes, and the regulation of Ig-gene transcription. The chapter concludes with the application of our knowledge of the molecular V κ J κ J κ 3′ C κ VJ C κ Polyadenylation RNA splicing (A) n 5′ L L a73 Genetic Model Compatible with Ig Structure a73 Multigene Organization of Ig Genes a73 Variable-Region Gene Rearrangements a73 Mechanism of Variable-Region DNA Rearrangements a73 Generation of Antibody Diversity a73 Class Switching among Constant-Region Genes a73 Expression of Ig Genes a73 Synthesis, Assembly, and Secretion of Immunoglobulins a73 Regulation of Ig-Gene Transcription a73 Antibody Genes and Antibody Engineering Organization and Expression of Immunoglobulin Genes O ?? ?? ??? ???? ?????????? ???????? ?? the vertebrate immune system is its ability to respond to an apparently limitless array of for- eign antigens. As immunoglobulin (Ig) sequence data accu- mulated, virtually every antibody molecule studied was found to contain a unique amino acid sequence in its vari- able region but only one of a limited number of invariant se- quences in its constant region. The genetic basis for this combination of constancy and tremendous variation in a single protein molecule lies in the organization of the im- munoglobulin genes. In germ-line DNA, multiple gene segments encode por- tions of a single immunoglobulin heavy or light chain. These gene segments are carried in the germ cells but cannot be transcribed and translated into complete chains until they are rearranged into functional genes. During B-cell matura- tion in the bone marrow, certain of these gene segments are randomly shuffled by a dynamic genetic system capable of generating more than 10 6 combinations. Subsequent processes increase the diversity of the repertoire of antibody binding sites to a very large number that exceeds 10 6 by at least two or three orders of magnitude. The processes of B- cell development are carefully regulated: the maturation of a progenitor B cell progresses through an ordered sequence of Ig-gene rearrangements, coupled with modifications to the gene that contribute to the diversity of the final product. By the end of this process, a mature, immunocompetent B cell will contain coding sequences for one functional heavy- chain variable-region and one light-chain variable-region. The individual B cell is thus antigenically committed to a specific epitope. After antigenic stimulation of a mature B cell in peripheral lymphoid organs, further rearrangement of constant-region gene segments can generate changes in the isotype expressed, which produce changes in the biolog- ical effector functions of the immunoglobulin molecule without changing its specificity. Thus, mature B cells contain chromosomal DNA that is no longer identical to germ-line Kappa Light-Chain Gene Rearrangement 8536d_ch05_105-136 8/22/02 2:46 PM Page 105 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: biology of immunoglobulin genes to the engineering of anti- body molecules for therapeutic and research applications. Chapter 11 covers in detail the entire process of B-cell devel- opment from the first gene rearrangements in progenitor B cells to final differentiation into memory B cells and anti- body-secreting plasma cells. Figure 5-1 outlines the sequen- tial stages in B-cell development, many of which result from critical rearrangements. Genetic Model Compatible with Ig Structure The results of the immunoglobulin-sequencing studies de- scribed in Chapter 4 revealed a number of features of immunoglobulin structure that were difficult to reconcile with classic genetic models. Any viable model of the 106 PART II Generation of B-Cell and T-Cell Responses VISUALIZING CONCEPTS FIGURE 5-1 Overview of B-cell development. The events that occur during maturation in the bone marrow do not require anti- gen, whereas activation and differentiation of mature B cells in pe- ripheral lymphoid organs require antigen. The labels mIgM and mIgD refer to membrane-associated Igs. IgG, IgA, and IgE are se- creted immunoglobulins. Lymphoid cell Partial heavy-chain gene rearrangement Hematopoietic stem cell Pro-B cell Complete heavy-chain gene rearrangement Pre-B cell Light-chain gene rearrangement Immature B cell Change in RNA processing Peripheral lymphoid organs Bone marrow Mature B cell Antigen stimulation Activated B cell Differentiation IgM-secreting plasma cells Class switching Memory B cells of various isotypes Plasma cells secreting various isotypes None Ig EXPRESSEDCELL None None μ Heavy chain + surrogate light chain mIgM mIgM + mIgD IgM IgG IgA IgE 8536d_ch05_105-136 8/22/02 2:46 PM Page 106 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: immunoglobulin genes had to account for the following properties of antibodies: a73 The vast diversity of antibody specificities a73 The presence in Ig heavy and light chains of a variable region at the amino-terminal end and a constant region at the carboxyl-terminal end a73 The existence of isotypes with the same antigenic specificity, which result from the association of a given variable region with different heavy-chain constant regions Germ-Line and Somatic-Variation Models Contended To Explain Antibody Diversity For several decades, immunologists sought to imagine a ge- netic mechanism that could explain the tremendous diversity of antibody structure. Two different sets of theories emerged. The germ-line theories maintained that the genome con- tributed by the germ cells, egg and sperm, contains a large repertoire of immunoglobulin genes; thus, these theories in- voked no special genetic mechanisms to account for anti- body diversity. They argued that the immense survival value of the immune system justified the dedication of a significant fraction of the genome to the coding of antibodies. In con- trast, the somatic-variation theories maintained that the genome contains a relatively small number of immunoglob- ulin genes, from which a large number of antibody specifici- ties are generated in the somatic cells by mutation or recombination. As the amino acid sequences of more and more im- munoglobulins were determined, it became clear that there must be mechanisms not only for generating antibody diver- sity but also for maintaining constancy. Whether diversity was generated by germ-line or by somatic mechanisms, a paradox remained: How could stability be maintained in the constant (C) region while some kind of diversifying mecha- nism generated the variable (V) region? Neither the germ-line nor the somatic-variation propo- nents could offer a reasonable explanation for this central feature of immunoglobulin structure. Germ-line proponents found it difficult to account for an evolutionary mechanism that could generate diversity in the variable part of the many heavy- and light-chain genes while preserving the constant region of each unchanged. Somatic-variation proponents found it difficult to conceive of a mechanism that could di- versify the variable region of a single heavy- or light-chain gene in the somatic cells without allowing alteration in the amino acid sequence encoded by the constant region. A third structural feature requiring an explanation emerged when amino acid sequencing of the human myeloma protein called Ti1 revealed that identical variable- region sequences were associated with both H9253 and H9262 heavy- chain constant regions. A similar phenomenon was observed in rabbits by C. Todd, who found that a particular allotypic marker in the heavy-chain variable region could be associ- ated with H9251, H9253, and H9262 heavy-chain constant regions. Consid- erable additional evidence has confirmed that a single variable-region sequence, defining a particular antigenic specificity, can be associated with multiple heavy-chain constant-region sequences; in other words, different classes, or isotypes, of antibody (e.g., IgG, IgM) can be expressed with identical variable-region sequences. Dreyer and Bennett Proposed the Two-Gene Model In an attempt to develop a genetic model consistent with the known findings about the structure of immunoglobulins, W. Dreyer and J. Bennett suggested, in their classic theoretical paper of 1965, that two separate genes encode a single im- munoglobulin heavy or light chain, one gene for the V region (variable region) and the other for the C region (constant re- gion). They suggested that these two genes must somehow come together at the DNA level to form a continuous mes- sage that can be transcribed and translated into a single Ig heavy or light chain. Moreover, they proposed that hundreds or thousands of V-region genes were carried in the germ line, whereas only single copies of C-region class and subclass genes need exist. The strength of this type of recombinational model (which combined elements of the germ-line and somatic- variation theories) was that it could account for those im- munoglobulins in which a single V region was combined with various C regions. By postulating a single constant- region gene for each immunoglobulin class and subclass, the model also could account for the conservation of necessary biological effector functions while allowing for evolutionary diversification of variable-region genes. At first, support for the Dreyer and Bennett hypothesis was indirect. Early studies of DNA hybridization kinetics us- ing a radioactive constant-region DNA probe indicated that the probe hybridized with only one or two genes, confirming the model’s prediction that only one or two copies of each constant-region class and subclass gene existed. However, in- direct evidence was not enough to overcome stubborn resis- tance in the scientific community to the hypothesis of Dreyer and Bennet. The suggestion that two genes encoded a single polypeptide contradicted the existing one gene–one polypeptide principle and was without precedent in any known biological system. As so often is the case in science, theoretical and intellec- tual understanding of Ig-gene organization progressed ahead of the available methodology. Although the Dreyer and Ben- nett model provided a theoretical framework for reconciling the dilemma between Ig-sequence data and gene organiza- tion, actual validation of their hypothesis had to wait for sev- eral major technological advances in the field of molecular biology. Organization and Expression of Immunoglobulin Genes CHAPTER 5 107 8536d_ch05_105-136 8/1/02 8:53 AM Page 107 mac79 Mac 79:45_BW:Goldsby et al. / Immunology 5e: Tonegawa’s Bombshell—Immunoglobulin Genes Rearrange In 1976, S. Tonegawa and N. Hozumi found the first direct evidence that separate genes encode the V and C regions of immunoglobulins and that the genes are rearranged in the course of B-cell differentiation. This work changed the field of immunology. In 1987, Tonegawa was awarded the Nobel Prize for this work. Selecting DNA from embryonic cells and adult myeloma cells—cells at widely different stages of development— Tonegawa and Hozumi used various restriction endonucle- ases to generate DNA fragments. The fragments were then separated by size and analyzed for their ability to hybridize with a radiolabeled mRNA probe. Two separate restriction fragments from the embryonic DNA hybridized with the mRNA, whereas only a single restriction fragment of the adult myeloma DNA hybridized with the same probe. Tone- gawa and Hozumi suggested that, during differentiation of lymphocytes from the embryonic state to the fully differenti- ated plasma-cell stage (represented in their system by the myeloma cells), the V and C genes undergo rearrangement. In the embryo, the V and C genes are separated by a large DNA segment that contains a restriction-endonuclease site; during differentiation, the V and C genes are brought closer together and the intervening DNA sequence is eliminated. The pioneering experiments of Tonegawa and Hozumi employed a tedious and time-consuming procedure that has since been replaced by the much more powerful approach of Southern-blot analysis. This method, now universally used to investigate the rearrangement of immunoglobulin genes, eliminates the need to elute the separated DNA restriction fragments from gel slices prior to analysis by hybridization with an immunoglobulin gene segment probe. Figure 5-2 shows the detection of rearrangement at the H9260 light-chain lo- cus by comparing the fragments produced by digestion of DNA from a clone of B-lineage cells with the pattern ob- tained by digestion of non-B cells (e.g., sperm or liver cells). The rearrangement of a V gene deletes an extensive section of germ-line DNA, thereby creating differences between re- arranged and unrearranged Ig loci in the distribution and number of restriction sites. This results in the generation of 108 PART II Generation of B-Cell and T-Cell Responses FIGURE 5-2 Experimental basis for diagnosis of rearrangement at an immunoglobulin locus. The number and size of restriction frag- ments generated by the treatment of DNA with a restriction enzyme is determined by the sequence of the DNA.The digestion of re- arranged DNA with a restriction enzyme (RE) yields a pattern of re- striction fragments that differ from those obtained by digestion of an unrearranged locus with the same RE. Typically, the fragments are an- alyzed by the technique of Southern blotting. In this example, a probe that includes a J gene segment is used to identify RE digestion frag- ments that include all or portions of this segment. As shown, re- arrangement results in the deletion of a segment of germ-line DNA and the loss of the restriction sites that it includes. It also results in the joining of gene segments, in this case a V and a J segment, that are separated in the germ line. Consequently, fragments dependent on the presence of this segment for their generation are absent from the restriction-enzyme digest of DNA from the rearranged locus. Fur- thermore, rearranged DNA gives rise to novel fragments that are ab- sent from digests of DNA in the germ-line configuration. This can be useful because both B cells and non-B cells have two immunoglobu- lin loci. One of these is rearranged and the other is not. Consequently, unless a genetic accident has resulted in the loss of the germ-line lo- cus, digestion of DNA from a myeloma or normal B-cell clone will produce a pattern of restriction that includes all of those in a germ- line digest plus any novel fragments that are generated from the change in DNA sequence that accompanies rearrangement. Note that only one of the several J gene segements present is shown. 3′5′ V n RE V 2 RE J C V 1 RE RE RE 3′5′ V n RE V 2 RE V 1 RE RE RE Germ line Rearranged Germ line Rearranged Deleted Rearrangement J C Probe Probe RE digestion RE digestion Southern blot 8536d_ch05_105-136 8/22/02 2:46 PM Page 108 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: different restriction patterns by rearranged and unre- arranged loci. Extensive application of this approach has demonstrated that the Dreyer and Bennett two-gene model—one gene encoding the variable region and another encoding the constant region—applied to both heavy and light-chain genes. Multigene Organization of Ig Genes As cloning and sequencing of the light- and heavy-chain DNA was accomplished, even greater complexity was re- vealed than had been predicted by Dreyer and Bennett. The H9260 and H9261 light chains and the heavy chains are encoded by sepa- rate multigene families situated on different chromosomes (Table 5-1). In germ-line DNA, each of these multigene fam- ilies contains several coding sequences, called gene seg- ments, separated by noncoding regions. During B-cell maturation, these gene segments are rearranged and brought together to form functional immunoglobulin genes. Each Multigene Family Has Distinct Features The H9260 and H9261 light-chain families contain V, J, and C gene seg- ments; the rearranged VJ segments encode the variable re- gion of the light chains. The heavy-chain family contains V, D, J, and C gene segments; the rearranged VDJ gene seg- ments encode the variable region of the heavy chain. In each gene family, C gene segments encode the constant regions. Each V gene segment is preceded at its 5H11032 end by a small exon that encodes a short signal or leader (L) peptide that guides the heavy or light chain through the endoplasmic reticulum. The signal peptide is cleaved from the nascent light and heavy chains before assembly of the finished immunoglobulin mol- ecule. Thus, amino acids encoded by this leader sequence do not appear in the immunoglobulin molecule. H9261-CHAIN MULTIGENE FAMILY The first evidence that the light-chain variable region was ac- tually encoded by two gene segments appeared when Tone- gawa cloned the germ-line DNA that encodes the variable region of mouse H9261 light chain and determined its complete nucleotide sequence. When the nucleotide sequence was compared with the known amino acid sequence of the H9261- chain variable region, an unusual discrepancy was observed. Although the first 97 amino acids of the H9261-chain variable re- gion corresponded to the nucleotide codon sequence, the re- maining 13 carboxyl-terminal amino acids of the protein’s variable region did not. It turned out that many base pairs away a separate, 39-bp gene segment, called J for joining, en- coded the remaining 13 amino acids of the H9261-chain variable region. Thus, a functional H9261 variable-region gene contains two coding segments—a 5H11032 V segment and a 3H11032 J segment— which are separated by a noncoding DNA sequence in unre- arranged germ-line DNA. The H9261 multigene family in the mouse germ line contains three V H9261 gene segments, four J H9261 gene segments, and four C H9261 gene segments (Figure 5-3a). The J H9261 4 is a pseudogene, a de- fective gene that is incapable of encoding protein; such genes are indicated with the psi symbol (H9274). Interestingly, J H9261 4’s constant region partner, C H9261 4, is a perfectly functional gene. The V H9261 and the three functional J H9261 gene segments en- code the variable region of the light chain, and each of the three functional C H9261 gene segments encodes the constant re- gion of one of the three H9261-chain subtypes (H92611, H92612, and H92613). In humans, the lambda locus is more complex. There are 31 functional V H9261 gene segments, 4 J H9261 segments, and 7 C H9261 segments. In additional to the functional gene seg- ments, the human lambda complex contains many V H9261 ,J H9261 , and C H9261 pseudogenes. H9260-CHAIN MULTIGENE FAMILY The H9260-chain multigene family in the mouse contains approx- imately 85 V H9260 gene segments, each with an adjacent leader se- quence a short distance upstream (i.e., on the 5H11032 side). There are five J H9260 gene segments (one of which is a nonfunctional pseudogene) and a single C H9260 gene segment (Figure 5-3b). As in the H9261 multigene family, the V H9260 and J H9260 gene segments en- code the variable region of the H9260 light chain, and the C H9260 gene segment encodes the constant region. Since there is only one C H9260 gene segment, there are no subtypes of H9260 light chains. Comparison of parts a and b of Figure 5-3 shows that the arrangement of the gene segments is quite different in the H9260 and H9261 gene families. The H9260-chain multigene family in hu- mans, which has an organization similar to that of the mouse, contains approximately 40 V H9260 gene segments, 5 J H9260 segments, and a single C H9260 segment. HEAVY-CHAIN MULTIGENE FAMILY The organization of the immunoglobulin heavy-chain genes is similar to, but more complex than, that of the H9260 and H9261 light-chain genes (Figure 5-3c). An additional gene segment encodes part of the heavy-chain variable region. The existence of this gene segment was first proposed by Leroy Hood and his colleagues, who compared the heavy-chain variable-region amino acid sequence with the V H and J H nucleotide sequences. The V H gene segment was found to encode amino acids 1 to 94 and the J H gene segment Organization and Expression of Immunoglobulin Genes CHAPTER 5 109 TABLE 5-1 Chromosomal locations of immunoglobulin genes in human and mouse CHROMOSOME Gene Human Mouse H9261 Light chain 22 16 H9260 Light chain 26 Heavy chain 14 12 8536d_ch05_105-136 8/22/02 2:46 PM Page 109 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: was found to encode amino acids 98 to 113; however, neither of these gene segments carried the information to encode amino acids 95 to 97. When the nucleotide sequence was de- termined for a rearranged myeloma DNA and compared with the germ-line DNA sequence, an additional nucleotide sequence was observed between the V H and J H gene seg- ments. This nucleotide sequence corresponded to amino acids 95 to 97 of the heavy chain. From these results, Hood and his colleagues proposed that a third germ-line gene segment must join with the V H and J H gene segments to encode the entire variable region of the heavy chain. This gene segment, which encoded amino acids within the third complementarity-determining region (CDR3), was designated D for diversity, because of its contri- bution to the generation of antibody diversity. Tonegawa and his colleagues located the D gene segments within mouse germ-line DNA with a cDNA probe complementary to the D region, which hybridized with a stretch of DNA lying be- tween the V H and J H gene segments. The heavy-chain multigene family on human chromo- some 14 has been shown by direct sequencing of DNA to contain 51 V H gene segments located upstream from a clus- ter of 27 functional D H gene segments. As with the light- chain genes, each V H gene segment is preceded by a leader sequence a short distance upstream. Downstream from the D H gene segments are six functional J H gene segments, fol- lowed by a series of C H gene segments. Each C H gene seg- ment encodes the constant region of an immunoglobulin heavy-chain isotype. The C H gene segments consist of coding exons and noncoding introns. Each exon encodes a separate domain of the heavy-chain constant region. A similar heavy- chain gene organization is found in the mouse. The conservation of important biological effector func- tions of the antibody molecule is maintained by the limited number of heavy-chain constant-region genes. In humans and mice, the C H gene segments are arranged sequentially in the order C H9262 ,C H9254 ,C H9253 ,C H9280 ,C H9251 (see Figure 5-3c). This sequential arrangement is no accident; it is generally related to the se- quential expression of the immunoglobulin classes in the course of B-cell development and the initial IgM response of a B cell to its first encounter with an antigen. Variable-Region Gene Rearrangements The preceding sections have shown that functional genes that encode immunoglobulin light and heavy chains are 110 PART II Generation of B-Cell and T-Cell Responses VISUALIZING CONCEPTS 3′5′ 1.3 kb 1.7 kb 1.4 kb 19 kb 1.3 kb 2.0 kb 1.2 kb 70 kb V λ 2 J λ 2C λ 2J λ 4C λ 4V λ 1J λ 3C λ 3J λ 1C λ 1L L ψ (a) λ-chain DNA 3′5′ 2.5 kb 23 kb V κ n J κ C κ (b) κ-chain DNA n = ～85 ψ V κ 1LV κ 2L L 3′5′ 34 kb 55 kb 4.5 kb 6.5 kb V H 1C μ C γ 3 (c) Heavy-chain DNA n = ～134 V H n D H 13D H 1J H 1C δ J H 4C γ 1C γ 2b C γ 2a C ε C α 21 kb 15 kb 14 kb 12 kb L L FIGURE 5-3 Organization of immunoglobulin germ-line gene segments in the mouse: (a) H9261 light chain, (b) H9260 light chain, and (c) heavy chain. The H9261 and H9260 light chains are encoded by V, J, and C gene segments. The heavy chain is encoded by V, D, J, and C gene segments. The distances in kilobases (kb) separating the various gene segments in mouse germ-line DNA are shown below each chain diagram. 8536d_ch05_105-136 8/22/02 2:46 PM Page 110 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: assembled by recombinational events at the DNA level. These events and the parallel events involving T-receptor genes are the only known site-specific DNA rearrangements in verte- brates. Variable-region gene rearrangements occur in an or- dered sequence during B-cell maturation in the bone marrow. The heavy-chain variable-region genes rearrange first, then the light-chain variable-region genes. At the end of this process, each B cell contains a single functional variable- region DNA sequence for its heavy chain and another for its light chain. The process of variable-region gene rearrangement pro- duces mature, immunocompetent B cells; each such cell is committed to produce antibody with a binding site encoded by the particular sequence of its rearranged V genes. As de- scribed later in this chapter, rearrangements of the heavy- chain constant-region genes will generate further changes in the immunoglobulin class (isotype) expressed by a B cell, but those changes will not affect the cell’s antigenic specificity. The steps in variable-region gene rearrangement occur in an ordered sequence, but they are random events that result in the random determination of B-cell specificity. The order, mechanism, and consequences of these rearrangements are described in this section. Light-Chain DNA Undergoes V-J Rearrangements Expression of both H9260 and H9261 light chains requires rearrange- ment of the variable-region V and J gene segments. In hu- mans, any of the functional V H9261 genes can combine with any of the four functional J H9261 -C H9261 combinations. In the mouse, things are slightly more complicated. DNA rearrangement can join the V H9261 1 gene segment with either the J H9261 1 or the J H9261 3 gene segment, or the V H9261 2 gene segment can be joined with the J H9261 2 gene segment. In human or mouse H9260 light-chain DNA, any one of the V H9260 gene segments can be joined with any one of the functional J H9261 gene segments. Rearranged H9260 and H9261 genes contain the following regions in order from the 5H11032 to 3H11032 end: a short leader (L) exon, a non- coding sequence (intron), a joined VJ gene segment, a second intron, and the constant region. Upstream from each leader gene segment is a promoter sequence. The rearranged light- chain sequence is transcribed by RNA polymerase from the L exon through the C segment to the stop signal, generating a light-chain primary RNA transcript (Figure 5-4). The in- trons in the primary transcript are removed by RNA- processing enzymes, and the resulting light-chain messenger Organization and Expression of Immunoglobulin Genes CHAPTER 5 111 FIGURE 5-4 Kappa light-chain gene rearrangement and RNA pro- cessing events required to generate a H9260 light-chain protein. In this example, rearrangement joins V H9260 23 and J H9260 4. Germ-line κ-chain DNA 5′ V κ 1 V κ 23 V κ n J κ C κ 3′ 3′ V κ J κ V κ J κ J κ J κ C κ Rearranged κ-chain DNA V-J joining 3′ C κ Transcription Primary RNA transcript VJ C κmRNA VJ C κ Nascent polypeptide VJ κ light chain C κ Polyadenylation RNA splicing Translation C κ V κ (A) n Poly-A tail ψ L 5′ V κ 1 5′ L L L L L L L 8536d_ch05_105-136 8/22/02 2:46 PM Page 111 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: RNA then exits from the nucleus. The light-chain mRNA binds to ribosomes and is translated into the light-chain pro- tein. The leader sequence at the amino terminus pulls the growing polypeptide chain into the lumen of the rough en- doplasmic reticulum and is then cleaved, so it is not present in the finished light-chain protein product. Heavy-Chain DNA Undergoes V-D-J Rearrangements Generation of a functional immunoglobulin heavy-chain gene requires two separate rearrangement events within the variable region. As illustrated in Figure 5-5, a D H gene seg- ment first joins to a J H segment; the resulting D H J H segment then moves next to and joins a V H segment to generate a V H D H J H unit that encodes the entire variable region. In heavy-chain DNA, variable-region rearrangement produces a rearranged gene consisting of the following sequences, starting from the 5H11032 end: a short L exon, an intron, a joined VDJ segment, another intron, and a series of C gene seg- ments. As with the light-chain genes, a promoter sequence is located a short distance upstream from each heavy-chain leader sequence. Once heavy-chain gene rearrangement is accomplished, RNA polymerase can bind to the promoter sequence and transcribe the entire heavy-chain gene, including the introns. Initially, both C H9262 and C H9254 gene segments are transcribed. Dif- ferential polyadenylation and RNA splicing remove the in- trons and process the primary transcript to generate mRNA including either the C H9262 or the C H9254 transcript. These two mRNAs are then translated, and the leader peptide of the re- sulting nascent polypeptide is cleaved, generating finished H9262 and H9254 chains. The production of two different heavy-chain mRNAs allows a mature, immunocompetent B cell to express both IgM and IgD with identical antigenic specificity on its surface. 112 PART II Generation of B-Cell and T-Cell Responses FIGURE 5-5 Heavy-chain gene rearrangement and RNA process- ing events required to generate finished H9262 or H9254 heavy-chain protein. Two DNA joinings are necessary to generate a functional heavy-chain gene: a D H to J H joining and a V H to D H J H joining. In this example, V H 21, D H 7, and J H 3 are joined. Expression of functional heavy-chain genes, although generally similar to expression of light-chain genes, involves differential RNA processing, which generates several differ- ent products, including H9262 or H9254 heavy chains. Each C gene is drawn as a single coding sequence; in reality, each is organized as a series of exons and introns. Primary RNA transcript mRNA Nascent polypeptide 5′ V H 1V H n D H 7D H 13D H 1 J HGerm-line H-chain DNA D-J joining 5′ V H 1V H 20 J H VDJ Rearranged H-chain DNA Transcription VJ Polyadenylation RNA splicing D μ heavy chain VJD VJD L L L L L L L 3′ C μ C γ 3C δ C γ 1C γ 2b C γ 2a C ε C α 3′ C μ C γ 3C δ C γ 1C γ 2b C γ 2a C ε C α C μ Translation (A) n C μ C μ VJDC δ Translation (A) n VJDC δ VJDC δ or or or δ heavy chain 3′ VJDC μ C δ 5′ L L L 3′5′ V H 1C μ C γ 3V H 21 D H 1D H 6D H J H C δ C γ 1C γ 2b C γ 2a C ε C α V-DJ joining V H nL L L 8536d_ch05_105-136 8/22/02 2:47 PM Page 112 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: Mechanism of Variable-Region DNA Rearrangements Now that we’ve seen the results of variable-region gene re- arrangements, let’s examine in detail how this process occurs during maturation of B cells. Recombination Signal Sequences Direct Recombination The discovery of two closely related conserved sequences in variable-region germ-line DNA paved the way to fuller un- derstanding of the mechanism of gene rearrangements. DNA sequencing studies revealed the presence of unique recombi- nation signal sequences (RSSs) flanking each germ-line V, D, and J gene segment. One RSS is located 3H11032 to each V gene segment, 5H11032 to each J gene segment, and on both sides of each D gene segment. These sequences function as signals for the recombination process that rearranges the genes. Each RSS contains a conserved palindromic heptamer and a conserved AT-rich nonamer sequence separated by an intervening se- quence of 12 or 23 base pairs (Figure 5-6a). The intervening 12- and 23-bp sequences correspond, respectively, to one and two turns of the DNA helix; for this reason the sequences are called one-turn recombination signal sequences and two- turn signal sequences. The V H9260 signal sequence has a one-turn spacer, and the J H9260 signal sequence has a two-turn spacer. In H9261 light-chain DNA, this order is reversed; that is, the V H9261 signal sequence has a two-turn spacer, and the J H9261 signal sequence has a one-turn spacer. In heavy-chain DNA, the signal sequences of the V H and J H gene segments have two-turn spacers, the signals on either side of the D H gene segment have one-turn spacers (Figure 5-6b). Signal sequences having a one-turn spacer can join only with sequences having a two-turn spacer (the so- called one-turn/two-turn joining rule). This joining rule en- sures, for example, that a V L segment joins only to a J L segment and not to another V L segment; the rule likewise en- sures that V H ,D H , and J H segments join in proper order and that segments of the same type do not join each other. Gene Segments Are Joined by Recombinases V-(D)-J recombination, which takes place at the junctions between RSSs and coding sequences, is catalyzed by enzymes collectively called V(D)J recombinase. Identification of the enzymes that catalyze recombination of V, D, and J gene segments began in the late 1980s and is still ongoing. In 1990 David Schatz, Marjorie Oettinger, and David Baltimore first reported the identification of two recombination-activating genes, designated RAG-1 and RAG-2, whose encoded proteins act synergistically and are re- quired to mediate V-(D)-J joining. The RAG-1 and RAG-2 pro- teins and the enzyme terminal deoxynucleotidyl transferase (TdT) are the only lymphoid-specific gene products that have been shown to be involved in V-(D)-J rearrangement. The recombination of variable-region gene segments consists of the following steps, catalyzed by a system of re- combinase enzymes (Figure 5-7): a73 Recognition of recombination signal sequences (RSSs) by recombinase enzymes, followed by synapsis in which Organization and Expression of Immunoglobulin Genes CHAPTER 5 113 FIGURE 5-6 Two conserved sequences in light-chain and heavy- chain DNA function as recombination signal sequences (RSSs). (a) Both signal sequences consist of a conserved palindromic hep- tamer and conserved AT-rich nonamer; these are separated by nonconserved spacers of 12 or 23 base pairs. (b) The two types of RSS—designated one-turn RSS and two-turn RSS—have charac- teristic locations within H9261-chain, H9260-chain, and heavy-chain germ- line DNA. During DNA rearrangement, gene segments adjacent to the one-turn RSS can join only with segments adjacent to the two- turn RSS. (a) Nucleotide sequence of RSSs CACAGTG GTGTCAC 23 bp 23 bp ACAAAAACC TGTTTTTGG Heptamer Nonamer Two-turn RSS 12 bp 12 bp Nonamer One-turn RSS Heptamer CACTGTG GTGACAC GGTTTTTGT CCAAAAACA (b) Location of RSSs in germ-line immunoglobulin DNA 3′5′ V H Heavy-chain DNA D H J H C H 3′5′ V κ κ-chain DNA J κ C κ 3′5′ V λ λ-chain DNA J λ C λ L L L 8536d_ch05_105-136 8/22/02 2:47 PM Page 113 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: two signal sequences and the adjacent coding sequences (gene segments) are brought into proximity a73 Cleavage of one strand of DNA by RAG-1 and RAG-2 at the junctures of the signal sequences and coding sequences a73 A reaction catalyzed by RAG-1 and RAG-2 in which the free 3H11032-OH group on the cut DNA strand attacks the phosphodiester bond linking the opposite strand to the signal sequence, simultaneously producing a hairpin structure at the cut end of the coding sequence and a flush, 5H11032-phosphorylated, double-strand break at the signal sequence a73 Cutting of the hairpin to generate sites for the addition of P-region nucleotides, followed by the trimming of a few nucleotides from the coding sequence by a single- strand endonuclease a73 Addition of up to 15 nucleotides, called N-region nucleotides, at the cut ends of the V, D, and J coding sequences of the heavy chain by the enzyme terminal deoxynucleotidyl transferase a73 Repair and ligation to join the coding sequences and to join the signal sequences, catalyzed by normal double- strand break repair (DSBR) enzymes Recombination results in the formation of a coding joint, falling between the coding sequences, and a signal joint, be- tween the RSSs. The transcriptional orientation of the gene segments to be joined determines the fate of the signal joint and intervening DNA. When the two gene segments are in the same transcriptional orientation, joining results in dele- tion of the signal joint and intervening DNA as a circular ex- cision product (Figure 5-8). Less frequently, the two gene segments have opposite orientations. In this case joining oc- curs by inversion of the DNA, resulting in the retention of 114 PART II Generation of B-Cell and T-Cell Responses (a) Deletional joining 3′5′ V κ J κ RSS 3′5′ V κ J κ (b) Inversional joining 3′ 5′ 3′ Recognition of RSSs by RAG-1/2 and synapsis V κ LJ κ Coding joint 5′ 3′ Signal joint Signal joint Coding joint Single-strand DNA cleavage by RAG-1/2 Hairpin formation and double-strand DNA break by RAG-1/2 Random cleavage of hairpin by endonuclease generates sites for the addition of P-nucleotides Optional addition to H-chain segments of N-nucleotides by TdT Repair and ligation of coding and signal sequences to form joints by DSBR enzymes 1 2 3 4 5 = Two-turn RSS = One-turn RSS L L + FIGURE 5-7 Model depicting the general process of recombina- tion of immunoglobulin gene segments is illustrated with V H9260 and J H9260 . (a) Deletional joining occurs when the gene segments to be joined have the same transcriptional orientation (indicated by horizontal blue arrows). This process yields two products: a rearranged VJ unit that includes the coding joint, and a circular excision product con- sisting of the recombination signal sequences (RSSs), signal joint, and intervening DNA. (b) Inversional joining occurs when the gene segments have opposite transcriptional orientations. In this case, the RSSs, signal joint, and intervening DNA are retained, and the orien- tation of one of the joined segments is inverted. In both types of re- combination, a few nucleotides may be deleted from or added to the cut ends of the coding sequences before they are rejoined. FIGURE 5-8 Circular DNA isolated from thymocytes in which the DNA encoding the chains of the T-cell receptor (TCR) undergoes re- arrangement in a process like that involving the immunoglobulin genes. Isolation of this circular excision product is direct evidence for the mechanism of deletional joining shown in Figure 5-7. [From K. Okazaki et al., 1987, Cell 49:477.] 8536d_ch05_105-136 8/22/02 2:47 PM Page 114 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: both the coding joint and the signal joint (and intervening DNA) on the chromosome. In the human H9260 locus, about half of the V H9260 gene segments are inverted with respect to J H9260 and their joining is thus by inversion. Ig-Gene Rearrangements May Be Productive or Nonproductive One of the striking features of gene-segment recombination is the diversity of the coding joints that are formed between any two gene segments. Although the double-strand DNA breaks that initiate V-(D)-J rearrangements are introduced precisely at the junctions of signal sequences and coding se- quences, the subsequent joining of the coding sequences is imprecise. Junctional diversity at the V-J and V-D-J coding joints is generated by a number of mechanisms: variation in cutting of the hairpin to generate P-nucleotides, variation in trimming of the coding sequences, variation in N-nucleotide addition, and flexibility in joining the coding sequences. The introduction of randomness in the joining process helps gen- erate antibody diversity by contributing to the hypervariabil- ity of the antigen-binding site. (This phenomenon is covered in more detail below in the section on generation of antibody diversity.) Another consequence of imprecise joining is that gene segments may be joined out of phase, so that the triplet read- ing frame for translation is not preserved. In such a nonpro- ductive rearrangement, the resulting VJ or VDJ unit is likely to contain numerous stop codons, which interrupt transla- tion (Figure 5-9). When gene segments are joined in phase, the reading frame is maintained. In such a productive re- arrangement, the resulting VJ or VDJ unit can be translated in its entirety, yielding a complete antibody. If one allele rearranges nonproductively, a B cell may still be able to rearrange the other allele productively. If an in- phase rearranged heavy-chain and light-chain gene are not produced, the B cell dies by apoptosis. It is estimated that only one in three attempts at V L -J L joining, and one in three subsequent attempts at V H -D H J H joining, are productive. As a result, less than 1/9 (11%) of the early-stage pre-B cells in the bone marrow progress to maturity and leave the bone marrow as mature immunocompetent B cells. Allelic Exclusion Ensures a Single Antigenic Specificity B cells, like all somatic cells, are diploid and contain both ma- ternal and paternal chromosomes. Even though a B cell is Organization and Expression of Immunoglobulin Genes CHAPTER 5 115 J κ V κ CACTGTG GTGGACTAGG GAGGATGCTCC CACAGTG RSS RSS 2 3 4 5 GAGGATGCGACTAGG Glu Asp Ala Thr Arg 1 GAGGATGGGACTAGG Glu Asp Gly Thr Arg GAGGATTGGACTAGG Glu Asp Trp Thr Arg Productive rearrangements 2 3 GAGGATGCGGACTAGG Glu Asp Ala Asp Stop GAGGTGGACTAGG Glu Val Asp Stop Nonproductive rearrangements 4 5 1 Joining flexibility FIGURE 5-9 Junctional flexibility in the joining of immunoglobulin gene segments is illustrated with V H9260 and J H9260 . In-phase joining (arrows 1, 2, and 3) generates a productive rearrangement, which can be translated into protein. Out-of-phase joining (arrows 4 and 5) leads to a nonproductive rearrangement that contains stop codons and is not translated into protein. κκ λλ HH κλH ** ** κλH Paternal chromosomes Gene rearrangement Maternal chromosomes Maternal H chain Maternal κ chain Maternal H chain Paternal λ chain FIGURE 5-10 Because of allelic exclusion, the immunoglobulin heavy- and light-chain genes of only one parental chromosome are expressed per cell. This process ensures that B cells possess a single antigenic specificity. The allele selected for rearrangement is chosen randomly. Thus the expressed immunoglobulin may contain one ma- ternal and one paternal chain or both chains may derive from only one parent. Only B cells and T cells exhibit allelic exclusion. Asterisks (?) indicate the expressed alleles. 8536d_ch05_105-136 8/22/02 2:47 PM Page 115 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: diploid, it expresses the rearranged heavy-chain genes from only one chromosome and the rearranged light-chain genes from only one chromosome. The process by which this is ac- complished, called allelic exclusion, ensures that functional B cells never contain more than one V H D H J H and one V L J L unit (Figure 5-10). This is, of course, essential for the antigenic specificity of the B cell, because the expression of both alleles would render the B cell multispecific. The phe- nomenon of allelic exclusion suggests that once a productive V H -D H -J H rearrangement and a productive V L -J L rearrange- ment have occurred, the recombination machinery is turned off, so that the heavy- and light-chain genes on the homolo- gous chromosomes are not expressed. G. D. Yancopoulos and F. W. Alt have proposed a model to account for allelic exclusion (Figure 5-11). They suggest that once a productive rearrangement is attained, its encoded protein is expressed and the presence of this protein acts as a signal to prevent further gene rearrangement. According to their model, the presence of H9262 heavy chains signals the 116 PART II Generation of B-Cell and T-Cell Responses FIGURE 5-11 Model to account for allelic exclusion. Heavy-chain genes rearrange first, and once a productive heavy-chain gene rearrangement occurs, the H9262 protein product prevents rearrange- ment of the other heavy-chain allele and initiates light-chain gene rearrangement. In the mouse, rearrangement of H9260 light-chain genes precedes rearrangement of the H9261 genes, as shown here. In humans, V κ J κ μ Productive allele #1D H J H V H D H J H V H D H J H V H D H J H VH D H J H Productive allele #2 Cell death Nonproductive allele #1 V κ J κ V λ J λ V λ J λ Productive allele #1 Productive allele #2 Productive allele #1 Productive allele #2 Nonproductive allele #1 Nonproductive allele #2 Nonproductive allele #2 Cell death Nonproductive allele #2 Nonproductive allele #1 μ + λ chains inhibit rearrangement of λ allele #2 μ + κ chains inhibit λ rearrangement μ + κ chains inhibit rearrangement of κ allele #2 and λ rearrangement μ heavy chain inhibits rearrangement of μ allele #2 and induces κ rearrangement Progenitor B cell Ig μ μ + κ μ μ + κ μ + λ μ + λ μ μ μ maturing B cell to turn off rearrangement of the other heavy-chain allele and to turn on rearrangement of the H9260 light-chain genes. If a productive H9260 rearrangement occurs, H9260 light chains are produced and then pair with H9262 heavy chains to form a complete antibody molecule. The presence of this antibody then turns off further light-chain rearrangement. If H9260 rearrangement is nonproductive for both H9260 alleles, re- arrangement of the H9261-chain genes begins. If neither H9261 allele rearranges productively, the B cell presumably ceases to ma- ture and soon dies by apoptosis. Two studies with transgenic mice have supported the hy- pothesis that the protein products encoded by rearranged heavy- and light-chain genes regulate rearrangement of the remaining alleles. In one study, transgenic mice carrying a rearranged H9262 heavy-chain transgene were prepared. The H9262 transgene product was expressed by a large percentage of the B cells, and rearrangement of the endogenous immunoglob- ulin heavy-chain genes was blocked. Similarly, cells from a transgenic mouse carrying a H9260 light-chain transgene did not either H9260 or H9261 rearrangement can proceed once a productive heavy- chain rearrangement has occurred. Formation of a complete immunoglobulin inhibits further light-chain gene rearrangement. If a nonproductive rearrangement occurs for one allele, then the cell attempts rearrangement of the other allele. [Adapted from G. D. Yancopoulos and F. W. Alt, 1986, Annu. Rev. Immunol. 4:339.] 8536d_ch05_105-136 8/22/02 2:47 PM Page 116 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: rearrange the endogenous H9260-chain genes when the H9260 trans- gene was expressed and was associated with a heavy chain to form complete immunoglobulin. These studies suggest that expression of the heavy- and light-chain proteins may indeed prevent gene rearrangement of the remaining alleles and thus account for allelic exclusion. Generation of Antibody Diversity As the organization of the immunoglobulin genes was deci- phered, the sources of the vast diversity in the variable region began to emerge. The germ-line theory, mentioned earlier, argued that the entire variable-region repertoire is encoded in the germ line of the organism and is transmitted from par- ent to offspring through the germ cells (egg and sperm). The somatic-variation theory held that the germ line contains a limited number of variable genes, which are diversified in the somatic cells by mutational or recombinational events dur- ing development of the immune system. With the cloning and sequencing of the immunoglobulin genes, both models were partly vindicated. To date, seven means of antibody diversification have been identified in mice and humans: a73 Multiple germ-line gene segments a73 Combinatorial V-(D)-J joining a73 Junctional flexibility a73 P-region nucleotide addition (P-addition) a73 N-region nucleotide addition (N-addition) a73 Somatic hypermutation a73 Combinatorial association of light and heavy chains Although the exact contribution of each of these avenues of diversification to total antibody diversity is not known, they each contribute significantly to the immense number of dis- tinct antibodies that the mammalian immune system is ca- pable of generating. There Are Numerous Germ-Line V, D, and J Gene Segments An inventory of functional V, D, and J gene segments in the germ-line DNA of one human reveals 51 V H , 25 D, 6 J H , 40 V H9260 ,5 J H9260 , 31 V H9261 , and 4 J H9261 gene segments. In addition to these functional segments, there are many pseudogenes. It should be borne in mind that these numbers were largely derived from a landmark study that sequenced the DNA of the immunoglobulin loci of a single individual. The immuno- globulin loci of other individuals might contain slightly dif- ferent numbers of particular types of gene segments. In the mouse, although the numbers are known with less precision than in the human, there appear to be about 85 V H9260 gene segments and 134 V H gene segments, 4 functional J H ,4 functional J H9260 , 3 functional J H9261 , and an estimated 13 D H gene segments, but only three V H9261 gene segments. Although the number of germ-line genes found in either humans or mice is far fewer than predicted by early proponents of the germ- line model, multiple germ-line V, D, and J gene segments clearly do contribute to the diversity of the antigen-binding sites in antibodies. Combinatorial V-J and V-D-J Joining Generates Diversity The contribution of multiple germ-line gene segments to an- tibody diversity is magnified by the random rearrangement of these segments in somatic cells. It is possible to calculate how much diversity can be achieved by gene rearrangments (Table 5-2). In humans, the ability of any of the 51 V H gene segments to combine with any of the 27 D H segments and any of the 6 J H segments allows a considerable amount of heavy-chain gene diversity to be generated (51 H11003 27 H110036 H11005 8262 possible combinations). Similarly, 40 V H9260 gene segments randomly combining with 5 J H9260 segments has the potential of generating 200 possible combinations at the H9260 locus, while 30 V H9261 and 4 J H9261 gene segments allow up to 120 possible combina- tions at the human H9261 locus. It is important to realize that these are minimal calculations of potential diversity. Junc- tional flexibility and P- and N-nucleotide addition, as men- tioned above, and, especially, somatic hypermutation, which will be described shortly, together make an enormous contri- bution to antibody diversity. Although it is not possible to make an exact calculation of their contribution, most work- ers in this field agree that they raise the potential for antibody combining-site diversity in humans to well over 10 10 . This does not mean that, at any given time, a single individual has a repertoire of 10 10 different antibody combining sites. These very large numbers describe the set of possible variations, of which any individual carries a subset that is smaller by several orders of magnitude. Junctional Flexibility Adds Diversity The enormous diversity generated by means of V, D, and J combinations is further augmented by a phenomenon called junctional flexibility. As described above, recombination in- volves both the joining of recombination signal sequences to form a signal joint and the joining of coding sequences to form a coding joint (see Figure 5-7). Although the signal se- quences are always joined precisely, joining of the coding se- quences is often imprecise. In one study, for example, joining of the V H9260 21 and J H9260 1 coding sequences was analyzed in several pre-B cell lines. Sequence analysis of the signal and coding joints revealed the contrast in junctional precision (Figure 5-12). As illustrated previously, junctional flexibility leads to many nonproductive rearrangements, but it also generates productive combinations that encode alternative amino acids at each coding joint (see Figure 5-9), thereby increasing antibody diversity. The amino acid sequence variation gener- Organization and Expression of Immunoglobulin Genes CHAPTER 5 117 8536d_ch05_105-136 8/1/02 8:53 AM Page 117 mac79 Mac 79:45_BW:Goldsby et al. / Immunology 5e: 118 PART II Generation of B-Cell and T-Cell Responses TABLE 5-2 Combinatorial antibody diversity in humans and mice LIGHT CHAINS Multiple germ-line segments Heavy chain H9260H9261 ESTIMATED NUMBER OF SEGMENTS IN HUMANS ? V 51 40 30 D 27 0 0 J 654 Combinatorial V-D-J and V-J joining (possible number of combinations) 51 H11003 27 H11003 6 H11005 8262 40 H11003 5 H11005 200 30 H11003 4 H11005 120 Possible combinatorial associations of heavy and light chains ? 8262 H11003 (200 H11003 120) H11005 2.64 H11003 10 6 ESTIMATED NUMBER OF SEGMENTS IN MICE ? V 134 85 2 D 13 0 0 J 443 Combinatorial V-D-J and V-J joining 134 H11003 13 H11003 4 H11005 6968 85 H11003 4 H11005 340 2 H11003 3 H11005 6 (possible number of combinations) Possible combinatorial associations 6968 H11003 (340 H11001 6) H11005 2.41 H11003 10 6 of heavy and light chains ? ? These numbers have been determined from studies of single subjects; slight differences may be seen among different individuals. Also, in the human case, only the functional gene segments have been listed. The genome contains additional segments that are incapable of rearrangement or contain stop codons or both. In the mouse case, the figures contained in the table are only best estimates, because the locus has not been completely sequenced. ? Because of the diversity contributed by junctional flexibility, P-region nucleotide addition, N-region nucleotide addition, and somatic mutation, the actual potential exceeds these estimates by several orders of magnitude. FIGURE 5-12 Experimental evidence for junctional flexibility in im- munoglobulin-gene rearrangement. The nucleotide sequences flank- ing the coding joints between V H9260 21 and J H9260 1 and the corresponding signal joint sequences were determined in four pre-B cell lines. The sequence constancy in the signal joints contrasts with the sequence variability in the coding joints. Pink and yellow shading indicate nu- cleotides derived from V H9260 21 and J H9260 1, respectively, and purple and or- ange shading indicate nucleotides from the two RSSs. 5′...CACTGTG 5′...GGATCCTCCC CACAGTG...3′ GTGGACGTT...3′ RSS J κ 1 V κ 21 RSS Cell line #1 Cell line #2 Cell line #3 Cell line #4 5′-GGATCC GGACGTT-3′ 5′-GGATC TGGACGTT-3′ 5′-GGATCCTC GTGGACGTT-3′ 5′-GGATCCT TGGACGTT-3′ 5′-CACTGTG CACAGTG-3′ 5′-CACTGTG CACAGTG-3′ 5′-CACTGTG CACAGTG-3′ 5′-CACTGTG CACAGTG-3′ Pre-B cell lines Coding joints (V κ 21 J κ 1) Signal joints (RSS/RSS) 8536d_ch05_105-136 8/22/02 2:48 PM Page 118 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: (Figure 5-13b). Evidence that TdT is responsible for the ad- dition of these N-nucleotides has come from transfection studies in fibroblasts. When fibroblasts were transfected with the RAG-1 and RAG-2 genes, V-D-J rearrangement occurred but no N-nucleotides were present in the coding joints. How- ever, when the fibroblasts were also transfected with the gene encoding TdT, then V-D-J rearrangement was accompanied by addition of N-nucleotides at the coding joints. Up to 15 N-nucleotides can be added to both the D H -J H and V H -D H J H joints. Thus, a complete heavy-chain variable region is encoded by a V H ND H NJ H unit. The additional heavy- chain diversity generated by N-region nucleotide addition is quite large because N regions appear to consist of wholly random sequences. Since this diversity occurs at V-D-J coding joints, it is localized in CDR3 of the heavy-chain genes. Somatic Hypermutation Adds Diversity in Already-Rearranged Gene Segments All the antibody diversity described so far stems from mech- anisms that operate during formation of specific variable regions by gene rearrangement. Additional antibody diver- sity is generated in rearranged variable-region gene units by a process called somatic hypermutation. As a result of so- matic hypermutation, individual nucleotides in VJ or VDJ units are replaced with alternatives, thus potentially altering the specificity of the encoded immunoglobulins. Normally, somatic hypermutation occurs only within germinal centers (see Chapter 11), structures that form in sec- ondary lymphoid organs within a week or so of immuniza- tion with an antigen that activates a T-cell-dependent B-cell response. Somatic hypermutation is targeted to rearranged V- regions located within a DNA sequence containing about 1500 nucleotides, which includes the whole of the VJ or VDJ segment. Somatic hypermutation occurs at a frequency ap- proaching 10 H110023 per base pair per generation. This rate is at least a hundred thousand-fold higher (hence the name hyper- mutation) than the spontaneous mutation rate, about 10 H110028 /bp/generation, in other genes. Since the combined length of the H-chain and L-chain variable-region genes is about 600 bp, one expects that somatic hypermutation will introduce at least one mutation per every two cell divisions in the pair of V H and V L genes that encode an antibody. The mechanism of somatic hypermutation has not yet been determined. Most of the mutations are nucleotide substitutions rather than deletions or insertions. Somatic hypermutation in- troduces these substitutions in a largely, but not completely, random fashion. Recent evidence suggests that certain nu- cleotide motifs and palindromic sequences within V H and V L may be especially susceptible to somatic hypermutation. Somatic hypermutations occur throughout the VJ or VDJ segment, but in mature B cells they are clustered within the CDRs of the V H and V L sequences, where they are most likely to influence the overall affinity for antigen. Following expo- sure to antigen, those B cells with higher-affinity receptors will be preferentially selected for survival. This result of this Organization and Expression of Immunoglobulin Genes CHAPTER 5 119 ated by junctional flexibility in the coding joints has been shown to fall within the third hypervariable region (CDR3) in immunoglobulin heavy-chain and light-chain DNA (Table 5-3). Since CDR3 often makes a major contribution to antigen binding by the antibody molecule, amino acid changes generated by junctional flexibility are important in the generation of antibody diversity. P-Addition Adds Diversity at Palindromic Sequences As described earlier, after the initial single-strand DNA cleav- age at the junction of a variable-region gene segment and at- tached signal sequence, the nucleotides at the end of the coding sequence turn back to form a hairpin structure (see Figure 5-7). This hairpin is later cleaved by an endonuclease. This second cleavage sometimes occurs at a position that leaves a short single strand at the end of the coding sequence. The subsequent addition of complementary nucleotides to this strand (P-addition) by repair enzymes generates a palin- dromic sequence in the coding joint, and so these nucleotides are called P-nucleotides (Figure 5-13a). Variation in the po- sition at which the hairpin is cut thus leads to variation in the sequence of the coding joint. N-Addition Adds Considerable Diversity by Addition of Nucleotides Variable-region coding joints in rearranged heavy-chain genes have been shown to contain short amino acid se- quences that are not encoded by the germ-line V, D, or J gene segments. These amino acids are encoded by nucleotides added during the D-J and V to D-J joining process by a ter- minal deoxynucleotidyl transferase (TdT) catalyzed reaction TABLE 5-3 Sources of sequence variation in complementarity-determining regions of immunoglobulin heavy- and light-chain genes Source of variation CDR1 CDR2 CDR3 Sequence V segment V segment V L -J L junction; encoded by: V H -D H -J H junctions Junctional H11002H11002 H11001 flexibility P-nucleotide H11002H11002 H11001 addition N-nucleotide H11002H11002 H11001 addition ? Somatic H11001H11001 H11001 hypermutation ? N-nucleotide addition occurs only in heavy-chain DNA. 8536d_ch05_105-136 8/22/02 2:48 PM Page 119 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: differential selection is an increase in the antigen affinity ofa population of B cells. The overall process, called affinity maturation, takes place within germinal centers, and is de- scribed more fully in Chapter 11. Claudia Berek and Cesar Milstein obtained experimental evidence demonstrating somatic hypermutation during the course of an immune response to a hapten-carrier conju- gate. These researchers were able to sequence mRNA that encoded antibodies raised against a hapten in response to primary, secondary, or tertiary immunization (first, second, or third exposure) with a hapten-carrier conjugate. The hapten they chose was 2-phenyl-5-oxazolone (phOx), cou- pled to a protein carrier. They chose this hapten because it had previously been shown that the majority of antibodies it induced were encoded by a single germ-line V H and V H9260 gene segment. Berek and Milstein immunized mice with the phOx-carrier conjugate and then used the mouse spleen cells to prepare hybridomas secreting monoclonal anti- bodies specific for the phOx hapten. The mRNA sequence for the H chain and H9260 light chain of each hybridoma was then determined to identify deviations from the germ-line sequences. The results of this experiment are depicted in Figure 5-14. Of the 12 hybridomas obtained from mice seven days after a primary immunization, all used a particular V H , the V H Ox-1 gene segment, and all but one used the same V L gene segment, V H9260 Ox-1. Moreover, only a few mutations from the germ-line sequence were present in these hybrido- mas. By day 14 after primary immunization, analysis of eight hybridomas revealed that six continued to use the germ-line V H Ox-1 gene segment and all continued to use the V H9260 Ox-1 gene segment. Now, however, all of these hybridomas included one or more mutations from the germ-line sequence. Hybridomas analyzed from the secondary and tertiary responses showed a larger percentage utilizing germ-line V H gene segments other than the V H Ox-1 gene. In those hybridoma clones that utilized the V H Ox-1 and V H9260 Ox-1 gene segments, most of the mutations were clustered in the CDR1 and CDR2 hypervariable regions. The number of mutations in the anti-phOx hybridomas progressively in- creased following primary, secondary, and tertiary immu- nizations, as did the overall affinity of the antibodies for phOx (see Figure 5-14). A Final Source of Diversity Is Combinatorial Association of Heavy and Light Chains In humans, there is the potential to generate 8262 heavy- chain genes and 320 light-chain genes as a result of variable- region gene rearrangements. Assuming that any one of the possible heavy-chain and light-chain genes can occur ran- domly in the same cell, the potential number of heavy- and light-chain combinations is 2,644,240. This number is prob- ably higher than the amount of combinatorial diversity actu- ally generated in an individual, because it is not likely that all V H and V L will pair with each other. Furthermore, the re- combination process is not completely random; not all V H , D, or V L gene segments are used at the same frequency. Some are used often, others only occasionally, and still others al- most never. Although the number of different antibody combining sites the immune system can generate is difficult to calculate with precision, we know that it is quite high. Because the very large number of new sequences created by junctional 120 PART II Generation of B-Cell and T-Cell Responses FIGURE 5-13 P-nucleotide and N-nucleotide addition during joining. (a) If cleavage of the hairpin intermediate yields a double- stranded end on the coding sequence, then P-nucleotide addition does not occur. In many cases, however, cleavage yields a single- stranded end. During subsequent repair, complementary nucleotides are added, called P-nucleotides, to produce palin- dromic sequences (indicated by brackets). In this example, four extra base pairs (blue) are present in the coding joint as the result of P-nucleotide addition. (b) Besides P-nucleotide addition, addi- tion of random N-nucleotides (light red) by a terminal deoxynu- cleotidyl transferase (TdT) can occur during joining of heavy-chain coding sequences. (a) P-nucleotide addition Hairpin Cleavage of hairpin generates sites for the addition of P-nucleotides Cleavage of hairpin generates sites for the addition of P-nucleotides TC AG AT TA Repair enzymes add complementary nucleotides TCGA TATA D J TCGTA AGCT A TA TATA D J D J (b) N-nucleotide addition Hairpin TC AG AT TA TdT adds N-nucleotides Repair enzymes add complementary nucleotides TCGA TATA D J TCGTA AGCT AGT TCA A TA TATA D J D J 8536d_ch05_105-136 8/22/02 2:48 PM Page 120 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: Organization and Expression of Immunoglobulin Genes CHAPTER 5 121 flexibility, P-nucleotide addition, and N-nucleotide addition are within the third CDR, they are positioned to influence the structure of the antibody binding site. In addition to these sources of antibody diversity, the phenomenon of so- matic hypermutation contributes enormously to the reper- toire after antigen stimulation. Class Switching among Constant-Region Genes After antigenic stimulation of a B cell, the heavy-chain DNA can undergo a further rearrangement in which the V H D H J H unit can combine with any C H gene segment. The exact mechanism of this process, called class switching or iso- type switching, is unclear, but it involves DNA flanking sequences (called switch regions) located 2–3 kb upstream from each C H segment (except C H9254 ). These switch regions, though rather large (2 to 10 kb), are composed of multiple copies of short repeats (GAGCT and TGGGG). One hy- pothesis is that a protein or system of proteins that consti- tute the switch recombinase recognize these repeats and upon binding carry out the DNA recombination that results in class switching. Intercellular regulatory proteins known as cytokines act as “switch factors” and play major roles in determining the particular immunoglobulin class that is ex- pressed as a consequence of switching. Interleukin 4 (IL-4), FIGURE 5-14 Experimental evidence for somatic mutation in vari- able regions of immunoglobulin genes. The diagram compares the mRNA sequences of heavy chains and of light chains from hybrido- mas specific for the phOx hapten. The horizontal solid lines repre- sent the germ-line V H and V H9260 Ox-1 sequences; dashed lines represent sequences derived from other germ-line genes. Blue shading shows the areas where mutations clustered; the blue circles with vertical lines indicate locations of mutations that encode a different amino acid than the germ-line sequence. These data show that the fre- quency of mutation (1) increases in the course of the primary re- sponse (day 7 vs. day 14) and (2) is higher after secondary and ter- tiary immunizations than after primary immunization. Moreover, the dissociation constant (K d ) of the anti-phOx antibodies decreases dur- ing the transition from the primary to tertiary response, indicating an increase in the overall affinity of the antibody. Note also that most of the mutations are clustered within CDR1 and CDR2 of both the heavy and the light chains. [Adapted from C. Berek and C. Milstein, 1987, Im- munol. Rev. 96:23.] Heavy–chain V regions Day 7 Primary γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ3 γ1 γ1 μ μ μ γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 γ1 CDR1 CDR3CDR2CDR1 CDR2 CDR3 J 3 3.7 (D) Hybridoma clone subclass Light–chain V regions K d × 10 –7 M J 5 Day 14 Secondary Tertiary 2.8 2.8 2.8 3.6 4.0 3.3 0.5 6.0 4.0 0.9 3.4 0.7 0.4 0.1 0.2 1.4 0.6 0.9 0.02 1.1 0.1 0.4 ≤ 0.02 1.0 ≤ 0.03 ≤ 0.03 ≤ 0.03 0.15 0.2 J 4 J 4 J 4 J 2 J 2 J 2 J 4 J 4 J 2 J 4 J 4 J 4 J 4 J 4 8536d_ch05_105-136 8/22/02 2:48 PM Page 121 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: for example, induces class switching from C H9262 to C H9253 1 or C H9280 . In some cases, IL-4 has been observed to induce class switch- ing in a successive manner: first from C H9262 to C H9253 1 and then from C H9253 1 to C H9280 (Figure 5-15). Examination of the DNA ex- cision products produced during class switching from C H9262 to C H9253 1 showed that a circular excision product containing C H9262 together with the 5H11032 end of the H92531 switch region (S H9253 1) and the 3H11032 end of the H9262 switch region (S H9262 ) was generated. Fur- thermore, the switch from C H9253 1 to C H9280 produced circular exci- sion products containing C H9253 1 together with portions of the H9262, H9253, and H9280 switch regions. Thus class switching depends upon the interplay of three elements: switch regions, a switch recombinase, and the cytokine signals that dictate the isotype to which the B cell switches. A more complete de- scription of the role of cytokines in class switching appears in Chapter 11. Expression of Ig Genes As in the expression of other genes, post-transcriptional processing of immunoglobulin primary transcripts is required to produce functional mRNAs (see Figures 5-4 and 5-5). The primary transcripts produced from re- arranged heavy-chain and light-chain genes contain inter- vening DNA sequences that include noncoding introns and J gene segments not lost during V-(D)-J rearrangement. In addition, as noted earlier, the heavy-chain C-gene 122 PART II Generation of B-Cell and T-Cell Responses FIGURE 5-15 Proposed mechanism for class switching induced by interleukin 4 in rearranged immunoglobulin heavy-chain genes. A switch site is located upstream from each C H segment except C H9254 . Identification of the indicated circular excision products containing portions of the switch sites suggested that IL-4 induces sequential class switching from C H9262 to C H9253 1 to C H9280 . 5′ 3′ C μ C γ 3VDJ C δ C γ 1C γ 2b S γ 2b C γ 2a S γ 2a C ε S ε C α S α 3′ S γ 1S γ 3S μ S γ 1 5′S γ 1 3′S γ 15′S μ 3′S μ Recombination at S μ and S γ 1 DNA looping S γ 3 C δ C γ 3 C μ S γ 3 C δ C γ 3 C μ 5′S ε 3′S γ 1 S γ 2a C γ 2a C γ 2b S γ 2b 5′ 3′ + + VDJ C γ 1C γ 2b S γ 2b C γ 2a S γ 2a C ε S ε C α S α DNA looping and recombination at S γ 1 and S ε 5′ 3′ VDJ C ε C α S α 5′ VDJ S μ C γ 1 C γ 1 C γ 2b S γ 2b C γ 2a S γ 2a C ε S ε C α S α L L L L 8536d_ch05_105-136 8/22/02 2:48 PM Page 122 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: Organization and Expression of Immunoglobulin Genes CHAPTER 5 123 segments are organized as a series of coding exons and noncoding introns. Each exon of a C H gene segment corre- sponds to a constant-region domain or a hinge region of the heavy-chain polypeptide. The primary transcript must be processed to remove the intervening DNA sequences, and the remaining exons must be connected by a process called RNA splicing. Short, moderately conserved splice sequences, or splice sites, which are located at the intron- exon boundaries within a primary transcript, signal the positions at which splicing occurs. Processing of the pri- mary transcript in the nucleus removes each of these in- tervening sequences to yield the final mRNA product. The mRNA is then exported from the nucleus to be translated by ribosomes into complete H or L chains. Heavy-Chain Primary Transcripts Undergo Differential RNA Processing Processing of an immunoglobulin heavy-chain primary transcript can yield several different mRNAs, which explains how a single B cell can produce secreted or membrane- bound forms of a particular immunoglobulin and simulta- neously express IgM and IgD. EXPRESSION OF MEMBRANE OR SECRETED IMMUNOGLOBULIN As explained in Chapter 4, a particular immunoglobulin can exist in either membrane-bound or secreted form. The two forms differ in the amino acid sequence of the heavy-chain carboxyl-terminal domains (C H 3/C H 3 in IgA, IgD, and IgG and C H 4/C H 4 in IgE and IgM). The secreted form has a hy- drophilic sequence of about 20 amino acids in the carboxyl- terminal domain; this is replaced in the membrane-bound form with a sequence of about 40 amino acids containing a hydrophilic segment that extends outside the cell, a hy- drophobic transmembrane segment, and a short hydrophilic segment at the carboxyl terminus that extends into the cyto- plasm (Figure 5-16a). For some time, the existence of these two forms seemed inconsistent with the structure of germ- line heavy-chain DNA, which had been shown to contain a single C H gene segment corresponding to each class and subclass. The resolution of this puzzle came from DNA sequenc- ing of the C H9262 gene segment, which consists of four exons (C H9262 1, C H9262 2, C H9262 3, and C H9262 4) corresponding to the four do- mains of the IgM molecule. The C H9262 4 exon contains a nu- cleotide sequence (called S) at its 3H11032 end that encodes the hydrophilic sequence in the C H 4 domain of secreted IgM. Two additional exons called M1 and M2 are located just 1.8 kb downstream from the 3H11032 end of the C H9262 4 exon. The M1 exon encodes the transmembrane segment, and M2 encodes the cytoplasmic segment of the C H 4 domain in membrane-bound IgM. Later DNA sequencing revealed that all the C H gene segments have two additional down- stream M1 and M2 exons that encode the transmembrane and cytoplasmic segments. The primary transcript produced by transcription of a re- arranged H9262 heavy-chain gene contains two polyadenylation signal sequences, or poly-A sites, in the C H9262 segment. Site 1 is located at the 3H11032 end of the C H9262 4 exon, and site 2 is at the 3H11032 end of the M2 exon (Figure 5-16b). If cleavage of the pri- mary transcript and addition of the poly-A tail occurs at site 1, the M1 and M2 exons are lost. Excision of the introns and splicing of the remaining exons then produces mRNA en- coding the secreted form of the heavy chain. If cleavage and polyadenylation of the primary transcript occurs instead at site 2, then a different pattern of splicing results. In this case, splicing removes the S sequence at the 3H11032 end of the C H9262 4 exon, which encodes the hydrophilic carboxyl-terminal end of the secreted form, and joins the remainder of the C H9262 4 exon with the M1 and M2 exons, producing mRNA for the membrane form of the heavy chain. Thus, differential processing of a common primary tran- script determines whether the secreted or membrane form of an immunoglobulin will be produced. As noted previ- ously, mature naive B cells produce only membrane-bound antibody, whereas differentiated plasma cells produce se- creted antibodies. It remains to be determined precisely how naive B cells and plasma cells direct RNA processing prefer- entially toward the production of mRNA encoding one form or the other. SIMULTANEOUS EXPRESSION OF IgM AND IgD Differential RNA processing also underlies the simultane- ous expression of membrane-bound IgM and IgD by ma- ture B cells. As mentioned already, transcription of rearranged heavy-chain genes in mature B cells produces primary transcripts containing both the C H9262 and C H9254 gene segments. The C H9262 and C H9254 , gene segments are close together in the rearranged gene (only about 5 kb apart), and the lack of a switch site between them permits the entire VDJC H9262 C H9254 region to be transcribed into a single primary RNA tran- script about 15 kb long, which contains four poly-A sites (Figure 5-17a). Sites 1 and 2 are associated with C H9262 ,as de- scribed in the previous section; sites 3 and 4 are located at similar places in the C H9254 gene segment. If the heavy-chain transcript is cleaved and polyadenylated at site 2 after the C H9262 exons, then the mRNA will encode the membrane form of the H9262 heavy chain (Figure 5-17b); if polyadenylation is instead further downstream at site 4, after the C H9254 exons, then RNA splicing will remove the intervening C H9262 exons and produce mRNA encoding the membrane form of the H9254 heavy chain (Figure 5-17c). Since the mature B cell expresses both IgM and IgD on its membrane, both processing pathways must occur si- multaneously. Likewise, cleavage and polyadenylation of the primary heavy-chain transcript at poly-A site 1 or 3 in 8536d_ch05_105-136 8/1/02 8:53 AM Page 123 mac79 Mac 79:45_BW:Goldsby et al. / Immunology 5e: 124 PART II Generation of B-Cell and T-Cell Responses FIGURE 5-16 Expression of secreted and membrane forms of the heavy chain by alternative RNA processing. (a) Amino acid sequence of the carboxyl-terminal end of secreted and membrane H9262 heavy chains. Residues are indicated by the single-letter amino acid code. Hydrophilic and hydrophobic residues and regions are indicated by purple and orange, respectively, and charged amino acids are indicated with a H11001 or H11002. The white regions of the (b) Primary H-chain transcript VDJ J (A) n Poly-A site 1 Poly-A site 2 C μ C δ Poly-A site 3 Poly-A site 4 Polyadenylation RNA transcript for secreted μ DJ μ1 μ2 μ3 μ4 μ1 μ2 μ3 μ4 S JVS (A) n RNA transcript for membrane μ DJ μ1 μ2 μ3 μ4JVS (A) n D μ1 μ2 μ3 μ4JVS (A) n D μ1 μ2 μ3 μ4JV RNA splicing mRNA encoding secreted μ chain mRNA encoding membrane μ chain M1 M2 M1 M2 M1 M2 Site 1 Site 2 L L L L L Outside (a) Key: Hydrophilic Hydrophobic T G V N A E E E G F N L F I V S A T L F S T T Y F L S V T L F K V K L L W T T Membrane Cytoplasm + – 556 563 575 576 SS bridge COOH C μ 4 556 Secreted μ CHO Membrane μ COOH COOH COOH 594 597 576576 568 556 556 568 E– E– – – – E– + + 597 594 T G K P T L Y N V S L I M S D T G G T C Y C μ 4 Encoded by S exon of C μ Encoded by M1 and M2 exons of C μ sequences are identical in both forms. (b) Structure of the pri- mary transcript of a rearranged heavy-chain gene showing the C H9262 exons and poly-A sites. Polyadenylation of the primary transcript at either site 1 or site 2 and subsequent splicing (indicated by V- shaped lines) generates mRNAs encoding either secreted or membrane H9262 chains. 8536d_ch05_105-136 8/22/02 3:07 PM Page 124 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: plasma cells and subsequent splicing will yield the secreted form of the H9262 or H9254 heavy chains, respectively (see Figure 5-16b). Synthesis, Assembly, and Secretion of Immunoglobulins Immunoglobulin heavy- and light-chain mRNAs are translated on separate polyribosomes of the rough endo- plasmic reticulum (RER). Newly synthesized chains con- tain an amino-terminal leader sequence, which serves to guide the chains into the lumen of the RER, where the sig- nal sequence is then cleaved. The assembly of light (L) and heavy (H) chains into the disulfide-linked and glycosylated immunoglobulin molecule occurs as the chains pass through the cisternae of the RER. The complete molecules are transported to the Golgi apparatus and then into secretory vesicles, which fuse with the plasma membrane (Figure 5-18). The order of chain assembly varies among the im- munoglobulin classes. In the case of IgM, the H and L chains assemble within the RER to form half-molecules, and then two half-molecules assemble to form the complete molecule. In the case of IgG, two H chains assemble, then an H 2 L inter- mediate is assembled, and finally the complete H 2 L 2 mole- cule is formed. Interchain disulfide bonds are formed, and the polypeptides are glycosylated as they move through the Golgi apparatus. If the molecule contains the transmembrane sequence of the membrane form, it becomes anchored in the membrane of a secretory vesicle and is inserted into the plasma mem- brane as the vesicle fuses with the plasma membrane (see Figure 5-18, insert). If the molecule contains the hydrophilic sequence of secreted immunoglobulins, it is transported as a free molecule in a secretory vesicle and is released from the cell when the vesicle fuses with the plasma membrane. Organization and Expression of Immunoglobulin Genes CHAPTER 5 125 FIGURE 5-17 Expression of membrane forms of H9262 and H9254 heavy chains by alternative RNA processing. (a) Structure of rearranged heavy-chain gene showing C H9262 and C H9254 exons and poly-A sites. (b) Structure of H9262 m transcript and H9262 m mRNA resulting from poly- adenylation at site 2 and splicing. (c) Structure of H9254 m transcript and H9254 m mRNA resulting from polyadenylation at site 4 and splicing. Both processing pathways can proceed in any given B cell. 3′ (a) H-chain primary transcript VDJ JM1 5′ M2μ1 μ2 μ3 μ4 Poly-A site 1 Poly-A site 2 C μ δ1 C δ M1M2 Poly-A site 4 Poly-A site 3 δ2 δ3 ～6.5 kb S SL (A) n JS (A) n μ1 μ2 μ3 μ4 Splicing M1 M2 M1 M2 μ m transcript 5′ VDJ μ m mRNA 5′ VDJ (b) Polyadenylation of primary transcript at site 2 μ m ← C μ L L (A) n Splicing M2 VDJ δ1 δ2 δ3 (c) Polyadenylation of primary transcript at site 4 ← δ m VDJ 5′ (A) n M1 δ m transcript δ m mRNA C μ C δ JM1M2μ1 μ2 μ3 μ4 δ1 δ2 δ3SSM1M2 L L 8536d_ch05_105-136 8/22/02 3:07 PM Page 125 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: Regulation of Ig-Gene Transcription The immunoglobulin genes are expressed only in B-lineage cells, and even within this lineage, the genes are expressed at different rates during different developmental stages. As with other eukaryotic genes, three major classes of cis regulatory sequences in DNA regulate transcription of immunoglobu- lin genes: a73 Promoters: relatively short nucleotide sequences, extending about 200 bp upstream from the transcription initiation site, that promote initiation of RNA transcription in a specific direction a73 Enhancers: nucleotide sequences situated some distance upstream or downstream from a gene that activate transcription from the promoter sequence in an orientation-independent manner a73 Silencers: nucleotide sequences that down-regulate transcription, operating in both directions over a distance. The locations of the three types of regulatory elements in germ-line immunoglobulin DNA are shown in Figure 5-19. All of these regulatory elements have clusters of sequence motifs that can bind specifically to one or more nuclear pro- teins. Each V H and V L gene segment has a promoter located just upstream from the leader sequence. In addition, the JH9260 cluster and each of the D H genes of the heavy-chain locus are preceded by promoters. Like other promoters, the immunoglobulin promoters contain a highly conserved AT- rich sequence called the TATA box, which serves as a site for the binding of a number of proteins that are necessary for the initiation of RNA transcription. The actual process of tran- scription is performed by RNA polymerase II, which starts transcribing DNA from the initiation site, located about 25 bp downstream of the TATA box. Ig promoters also contain an essential and conserved octamer that confers B-cell speci- ficity on the promoter. The octamer binds two transcription factors, oct-1, found in many cell types, and oct-2, found only in B cells. While much remains to be learned about the function of enhancers, they have binding sites for a number of proteins, many of which are transcription factors. A particularly im- portant role is played by two proteins encoded by the E2A gene which can undergo alternate splicing to generate two collaborating proteins, both of which bind to the H9262 and H9260 in- tronic enhancers. These proteins are essential for B-cell de- velopment and E2A knockout mice make normal numbers of T cells but show a total absence of B cells. Interestingly, trans- fection of these enhancer-binding proteins into a T cell line resulted in a dramatic increase in the transcription of H9262chain mRNA and even induced the T cell to undergo D H H11001 J H → D H J H rearrangement. Silencers may inhibit the activity of Ig 126 PART II Generation of B-Cell and T-Cell Responses Secreted Ig Secretory vesicles Heavy-chain translation Light-chain translation Leader Nascent Ig (leader cleaved) RER Cis Golgi Trans Golgi Trans Golgi reticulum Oligosaccharides Membrane Ig Fusion with membrane Secretory vesicle Transmembrane segment FIGURE 5-18 Synthesis, assembly, and secretion of the im- munoglobulin molecule. The heavy and light chains are synthesized on separate polyribosomes (polysomes). The assembly of the chains to form the disulfide-linked immunoglobulin molecule oc- curs as the chains pass through the cisternae of the rough endo- plasmic reticulum (RER) into the Golgi apparatus and then into secretory vesicles. The main figure depicts assembly of a secreted antibody. The inset depicts a membrane-bound antibody, which contains the carboxyl-terminal transmembrane segment. This form becomes anchored in the membrane of secretory vesicles and then is inserted into the cell membrane when the vesicles fuse with the membrane. 8536d_ch05_105-136 8/22/02 3:07 PM Page 126 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: enhancers in non-B cells. If so, they could be important con- tributors to the high levels of Ig gene transcription that are characteristic of B cells but absent in other cell types. One heavy-chain enhancer is located within the intron between the last (3H11032) J gene segment and the first (5H11032) C gene segment (C H9262 ), which encodes the H9262 heavy chain. Because this heavy-chain enhancer (E H9262 ) is located 5H11032 of the S H9262 switch site near C H9262 , it can continue to function after class switching has occurred. Another heavy-chain enhancer (3H11032 H9251 E) has been detected 3H11032 of the C H9251 gene segment. One H9260 light-chain enhancer (E H9260 ) is located between the J H9260 segment and the C H9260 segment, and another enhancer (3H11032 H9260 E) is located 3H11032 of the C H9260 segment. The H9261 light-chain enhancers are located 3H11032 of C H9261 4 and 3H11032 of C H9261 1. Silencers have been identified in heavy-chain and H9260-chain DNA, adjacent to enhancers, but not in H9261-chain DNA. DNA Rearrangement Greatly Accelerates Transcription The promoters associated with the immunoglobulin V gene segments bind RNA polymerase II very weakly, and the vari- able-region enhancers in germ-line DNA are quite distant from the promoters (about 250–300 kb), too remote to signif- icantly influence transcription. For this reason, the rate of transcription of V H and V L coding regions is negligible in un- rearranged germ-line DNA. Variable-region gene rearrange- ment brings a promoter and enhancer within 2 kb of each other, close enough for the enhancer to influence transcription from the nearby promoter. As a result, the rate of transcription of a rearranged V L J L or V H D H J H unit is as much as 10 4 times the rate of transcription of unrearranged V L or V H segments. This effect was demonstrated directly in a study in which B cells transfected with rearranged heavy-chain genes from which the enhancer had been deleted did not transcribe the genes, whereas B cells transfected with similar genes that con- tained the enhancer transcribed the transfected genes at a high rate. These findings highlight the importance of enhancers in the normal transcription of immunoglobulin genes. Genes that regulate cellular proliferation or prohibit cell death sometimes translocate to the immunoglobulin heavy- or light-chain loci. Here, under the influence of an im- munoglobulin enhancer, the expression of these genes is sig- nificantly elevated, resulting in high levels of growth promoting or cell death inhibiting proteins. Translocations of the c-myc and bcl-2 oncogenes have each been associated with malignant B-cell lymphomas. The translocation of c- myc leads to constitutive expression of c-Myc and an aggres- sive, highly proliferative B-cell lymphoma called Burkitt’s lymphoma. The translocation of bcl-2 leads to suspension of programmed cell death in B cells, resulting in follicular B-cell lymphoma. These cancer-promoting translocations are cov- ered in greater detail in Chapter 22. Ig-Gene Expression Is Inhibited in T Cells As noted earlier, germ-line DNA encoding the T-cell receptor (TCR) undergoes V-(D)-J rearrangement to generate func- tional TCR genes. Rearrangement of both immunoglobulin and TCR germ-line DNA occurs by similar recombination processes mediated by RAG-1 and RAG-2 and involving re- combination signal sequences with one-turn or two-turn spacers (see Figure 5-7). Despite the similarity of the processes, complete Ig-gene rearrangement of H and L chains occurs only in B cells and complete TCR-gene rearrangement is limited to T cells. Organization and Expression of Immunoglobulin Genes CHAPTER 5 127 FIGURE 5-19 Location of promoters (dark red), enhancers (green), and silencers (yellow) in mouse heavy-chain, H9260 light-chain, and H9261 light-chain germ-line DNA. Variable-region DNA rearrange- ment moves an enhancer close enough to a promoter that the en- hancer can activate transcription from the promoter. The promoters that precede the DH cluster, a number of the C genes and the J H9261 clus- ter are omitted from this diagram for the sake of clarity. 3′5′ V H C μ C γ 3L H-chain DNA V H D H C δ J H C γ 1C γ 2b C γ 2a C ε Key H11005 Promoter Enhancer Silencer C αPP E μ Silencers 3′ α E L E κ 3′ κ E 3′5′ V κ C κ κ-chain DNA PP V κ J κP V κ ψ Silencers L L L λ2–4E λ3–1E 3′5′ λ-chain DNA V λ 1C λ 3 P V λ 2C λ 2J λ 2J λ 4 C λ 1J λ 1J λ 3L P C λ 4 ψ L 8536d_ch05_105-136 8/22/02 3:07 PM Page 127 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: Hitoshi Sakano and coworkers have obtained results sug- gesting that a sequence within the H9260-chain 3H11032enhancer (3H11032 H9260 E) serves to regulate the joining of V H9260 to J H9260 in B and T cells. When a sequence known as the PU.1 binding site within the 3H11032H9260-chain enhancer was mutated, these researchers found that V H9260 -J H9260 joining occurred in T cells as well as B cells. They propose that binding of a protein expressed by T cells, but not B cells, to the unmutated H9260-chain enhancer normally prevents V H9260 -J H9260 joining in T cells. The identity of this DNA- binding protein in T cells remains to be determined. Similar processes may prevent rearrangement of heavy-chain and H9261- chain DNA in T cells. Antibody Genes and Antibody Engineering There are many clinical applications in which the exquisite specificity of a mouse monoclonal antibody would be useful. However, when mouse monoclonal antibodies are intro- duced into humans they are recognized as foreign and evoke an antibody response that quickly clears the mouse mono- clonal antibody from the bloodstream. In addition, circulat- ing complexes of mouse and human antibodies can cause allergic reactions. In some cases, the buildup of these com- plexes in organs such as the kidney can cause serious and even life-threatening reactions. Clearly, one way to avoid these undesirable reactions is to use human monoclonal an- tibodies for clinical applications. However, the preparation of human monoclonal antibodies has been hampered by nu- merous technical problems. In response to the difficulty of producing human monoclonal antibodies and the complica- tions resulting from the use of mouse monoclonal antibodies in humans, there is now a major effort to engineer mono- clonal antibodies and antibody binding sites with recombi- nant DNA technology. The growing knowledge of antibody gene structure and regulation has made possible what Cesar Milstein, one of the inventors of monoclonal antibody technology, has called “man-made antibodies.” It is now possible to design and con- struct genes that encode immunoglobulin molecules in which the variable regions come from one species and the constant regions come from another. New genes have been created that link nucleotide sequences coding nonantibody proteins with sequences that encode antibody variable re- gions specific for particular antigens. These molecular hy- brids or chimeras may be able to deliver powerful toxins to particular antigenic targets, such as tumor cells. Finally, by replacement of the immunoglobulin loci of one species with that of another, animals of one species have been endowed with the capacity to respond to immunization by producing antibodies encoded by the donor’s genetically transplanted Ig genes. By capturing a significant sample of all of the im- munoglobulin heavy- and light-chain variable-region genes via incorporation into libraries of bacteriophage, it has been possible to achieve significant and useful reconstructions of the entire antibody repertoires of individuals. The next few sections describe each of these types of antibody genetic en- gineering. Chimeric and Hybrid Monoclonal Antibodies Have Potent Clinical Potential One approach to engineering an antibody is to clone recom- binant DNA containing the promoter, leader, and variable- region sequences from a mouse antibody gene and the constant-region exons from a human antibody gene (Figure 5-20). The antibody encoded by such a recombinant gene is a mouse-human chimera, commonly known as a humanized antibody. Its antigenic specificity, which is determined by the variable region, is derived from the mouse DNA; its isotype, which is determined by the constant region, is derived from the human DNA. Because the constant regions of these chimeric antibodies are encoded by human genes, the anti- 128 PART II Generation of B-Cell and T-Cell Responses LIGHT-CHAIN GENES HEAVY-CHAIN GENES Promoter Mouse V L Human C L Mouse V H Human C H Plasmid DNA Transfected antibody-secreting myeloma cell Transfect into Ab – myeloma cells Chimeric mouse-human antibody Light-chain chimeric vector Heavy-chain chimeric vector Selection gene (amp R ) Ig Promoter FIGURE 5-20 Production of chimeric mouse-human monoclonal antibodies. Chimeric mouse-human heavy- and light-chain expres- sion vectors are produced. These vectors are transfected into Ab H11002 myeloma cells. Culture in ampicillin medium selects for transfected myeloma cells that secrete the chimeric antibody. [Adapted from M. Verhoeyen and L. Reichmann, 1988, BioEssays 8:74.] 8536d_ch05_105-136 8/22/02 3:07 PM Page 128 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: bodies have fewer mouse antigenic determinants and are far less immunogenic when administered to humans than mouse monoclonal antibodies (Figure 5-21a). The ability of the mouse variable regions remaining in these humanized antibodies to provide the appropriate binding site to allow specific recognition of the target antigen has encouraged fur- ther exploration of this approach. It is possible to produce chimeric human-mouse antibodies in which only the se- quences of the CDRs are of mouse origin (Figure 5-21b). An- other advantage of humanized chimeric antibodies is that they retain the biological effector functions of human anti- body and are more likely to trigger human complement acti- vation or Fc receptor binding. One such chimeric human- mouse antibody has been used to treat patients with B-cell varieties of non-Hodgkin’s lymphoma (see Clinical Focus). Chimeric monoclonal antibodies that function as im- munotoxins (see Figure 4-23) can also be prepared. In this case, the terminal constant-region domain in a tumor- specific monoclonal antibody is replaced with toxin chains (Figure 5-21c). Because these immunotoxins lack the terminal Fc domain, they are not able to bind to cells bearing Fc receptors. These immunotoxins can bind only to tumor cells, making them highly specific as therapeutic reagents. Heteroconjugates, or bispecific antibodies, are hybrids of two different antibody molecules (Figure 5-21d). They can be constructed by chemically crosslinking two different antibodies or by synthesizing them in hybridomas consist- ing of two different monoclonal-antibody-producing cell lines that have been fused. Both of these methods generate mixtures of monospecific and bispecific antibodies from which the desired bispecific molecule must be purified. Us- ing genetic engineering to construct genes that will encode molecules only with the two desired specificities is a much simpler and more elegant approach. Several bispecific mole- cules have been designed in which one half of the antibody has specificity for a tumor and the other half has specificity for a surface molecule on an immune effector cell, such as an NK cell, an activated macrophage, or a cytotoxic T lympho- cyte (CTL). Such heteroconjugates have been designed to activate the immune effector cell when it is crosslinked to the tumor cell so that it begins to mediate destruction of the tumor cell. Monoclonal Antibodies Can Be Constructed from Ig-Gene Libraries A quite different approach for generating monoclonal anti- bodies employs the polymerase chain reaction (PCR) to am- plify the DNA that encodes antibody heavy-chain and light-chain Fab fragments from hybridoma cells or plasma cells. A promoter region and EcoRI restriction site (see Chap- ter 23) are added to the amplified sequences, and the result- ing constructs are inserted into bacteriophage H9261, yielding separate heavy- and light-chain libraries. Cleavage with EcoRI and random joining of the heavy- and light-chain genes yield numerous novel heavy-light constructs (Figure 5-22). This procedure generates an enormous diversity of anti- body specificities—libraries with H1102210 10 unique members have been obtained—and clones containing these random combinations of H H11001 L chains can be rapidly screened for those secreting antibody to a particular antigen. The level of diversity is comparable to the human in vivo repertoire, and it is possible to demonstrate that specificities against a wide variety of antigens can be obtained from these libraries. Such a combinatorial library approach opens the possibility of ob- taining specific antibodies without any need whatsoever for immunization. However, the real challenge to bypassing in vivo immu- nization in the derivation of useful antibodies of high affin- ity lies in finding ways to mimic the biology of the humoral Organization and Expression of Immunoglobulin Genes CHAPTER 5 129 Mouse (c) Human Mouse monoclonal antibody (anti-tumor) Chimeric mouse-human antibody Chimeric immunotoxin Heteroconjugate Anti-T-cell receptor Anti-tumor antibody Toxin (a) (b) Mouse Human (d) Grafted CDRs FIGURE 5-21 Chimeric and hybrid monoclonal antibodies engi- neered by recombinant DNA technology. (a) Chimeric mouse-hu- man monoclonal antibody containing the V H and V L domains of a mouse monoclonal antibody (blue) and C L and C H domains of a hu- man monoclonal antibody (gray). (b) A chimeric monoclonal anti- body containing only the CDRs of a mouse monoclonal antibody (blue bands) grafted within the framework regions of a human mon- oclonal antibody is called a “humanized” monoclonal antibody. (c) A chimeric monoclonal antibody in which the terminal Fc domain is re- placed by toxin chains (white). (d) A heteroconjugate in which one- half of the mouse antibody molecule is specific for a tumor antigen and the other half is specific for the CD3/T-cell receptor complex. 8536d_ch05_105-136 8/22/02 3:07 PM Page 129 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: 130 PART II Generation of B-Cell and T-Cell Responses that will affect only the tumor cells and completely spare normal cells. If particu- lar types of cancer cells had antigens that were tumor specific, these antigens would be ideal targets for immune attack. Unfortunately, there are few such mole- cules known. However, a number of anti- gens are known that are restricted to the cell lineage in which the tumor originated and are expressed on the tumor cells. Many cell-lineage-specific antigens have been identified for B lymphocytes and B lymphomas, including immuno- globulin, the hallmark of the B cell, and CD20, a membrane-bound phosphopro- tein. CD20 has emerged as an attractive candidate for antibody-mediated im- munotherapy because it is present on B lymphomas, and antibody-mediated crosslinking does not cause it to down- regulate or internalize. Indeed, some years ago, mouse monoclonal antibodies were raised against CD20, and one of these has formed the basis for an anti-B- cell lymphoma immunotherapy. This ap- proach appears ready to take its place as an adjunct or alternative to radiation and chemotherapy. The development of this anti-tumor antibody is an excellent case study of the combined application of im- munological insights and molecular biol- ogy to engineer a novel therapeutic agent. The original anti-CD20 antibody was a mouse monoclonal antibody with murine H9253 heavy chains and H9260 light chains. The DNA sequences of the light- and heavy- chain variable regions of this antibody were amplified by PCR. Then a chimeric gene was created by replacing the CDR gene sequences of a human H92531 heavy chain with those from the murine heavy chain. In a similar maneuver, CDRs from the mouse H9260 were ligated into a human H9260 gene. The chimeric genes thus created were incorporated into vectors that per- mitted high levels of expression in mam- malian cells. When an appropriate cell line was co-transfected with both of these constructs, it produced chimeric antibod- ies containing CDRs of mouse origin to- gether with human variable-region frameworks and constant regions. After purification, the biological activity of the antibody was evaluated, first in vitro and then in a primate animal model. The initial results were quite promis- ing. The grafted human constant region supported effector functions such as the complement-mediated lysis or antibody- dependent cell-mediated cytotoxicity (ADCC) of human B lymphoid cells. Fur- thermore, weekly injections of the anti- body into monkeys resulted in the rapid and sustained depletion of B cells from pe- ripheral blood, lymph nodes, and even bone marrow. When the anti-CD20 anti- body infusions were stopped, the differen- tiation of new B cells from progenitor populations allowed B-cell populations eventually to recover and approach nor- mal levels. From these results, the hope grew that this immunologically active chimeric antibody could be used to clear entire B cell populations, including B lym- phoma cells, from the body in a way that spared other cell populations. This led to the trial of the antibody in human patients. The human trials enrolled patients with B-cell lymphoma who had a relapse after chemotherapy or radiation treat- ment. These trials addressed three im- portant issues: efficacy, safety, and immunogenicity. While not all patients re- sponded to treatment with anti-CD20, close to 50% exhibited full or partial re- mission. Thus, efficacy was demon- strated, because this level of response is comparable to the success rate with tradi- tional approaches that employ highly cyto- toxic drugs or radiation—it offers a truly alternative therapy. Side effects such as nausea, low blood pressure, and short- ness of breath were seen in some pa- tients (usually during or shortly after the initiation of therapy); these were, for the most part, not serious or life-threatening. Consequently, treatment with the Lymphomas are cancers of lymphatic tissue in which the tumor cells are of lymphocytic origin. There are two major forms of lymphoma: Hodgkin’s lymphoma and non-Hodgkin’s lymphoma. The less common form is Hodgkin’s lymphoma, named for its dis- coverer, Thomas Hodgkin, an English physician. This unusually gifted early pathologist, who worked without the ben- efit of a microscope, recognized this con- dition in several patients and first described the anatomical features of the disease in 1832. Because many tissue specimens taken from patients Hodgkin suspected of harboring the disease were saved in the Gordon Museum of Guy’s Hospital in London, it has been possible for later generations to judge the accu- racy of his diagnoses. Hodgkin has fared well. Studies of these preserved tissues confirm that he was right in about 60% of the cases, a surprising achievement, con- sidering the technology of the time. Actu- ally, most lymphoma is non-Hodgkin’s type and includes about 10 different types of disease. B-cell lymphomas are an im- portant fraction of these. For some years now, the major thera- pies directed against lymphomas have been radiation, chemotherapy, or a com- bination of both. While these therapies benefit large numbers of patients by in- creasing survival, relapses after treat- ment are common, and many treated patients experience debilitating side ef- fects. The side effects are an expected consequence of these therapies, because the agents used kill or severely damage a broad spectrum of normal cells as well as tumor cells. One of the holy grails of can- cer treatment is the discovery of therapies CLINICAL FOCUS Therapy for Non-Hodgkin’s Lymphoma and Other Diseases by Genetically Engineered Antibodies 8536d_ch05_105-136 8/22/02 3:07 PM Page 130 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: Organization and Expression of Immunoglobulin Genes CHAPTER 5 131 cancers that over-express a growth factor receptor called HER2 (human epidermal growth factor receptor 2). Many tumors that over-express HER2 grow faster and pose a more serious threat than those with normal levels of this protein on their sur- face. A chimeric anti-HER2 monoclonal antibody in which all of the protein except the CDRs are of human origin was created by genetic engineering. Specifically, the DNA sequences for the heavy-chain and light-chain CDRs were taken from cloned mouse genes encoding an anti-HER2 monoclonal antibody. As in the anti-CD20 strategy described above, each of the mouse CDR gene segments were used to replace the corresponding human CDR gene segments in human genes encoding the human IgG 1 heavy chain and the hu- man H9260 light chain. When this engineered antibody is used in combination with a chemotherapeutic drug, it is highly effec- tive against metastatic breast cancer. The effects on patients who were given only a chemotherapeutic drug were compared with those for patients receiving both the chemotherapeutic drug and the engi- neered anti-HER2 antibody. The combina- tion anti-HER2/chemotherapy treatment showed significantly reduced rates of tu- mor progression, a higher percentage of responding patients, and a higher one-year survival rate. Treatment with Herceptin, as this engineered monoclonal antibody is called, has become part of the standard repertoire of breast cancer therapies. The development of engineered and conventional monoclonal antibodies is one of the most active areas in the phar- maceutical industry. The table provides a partial compilation of monoclonal anti- bodies that have received approval from the Food and Drug Administration (FDA) for use in the treatment of human disease. Many more are in various stages of development and testing. chimeric anti-CD20 appears safe. Patients who received the antibody have been ob- served closely for the appearance of hu- man anti-mouse-Ig antibodies (HAMA) and for human anti-chimeric antibody (HACA) responses. Such responses were not observed. Therefore, the antibody was not immunogenic. The absence of such responses demonstrate that antibodies can be genetically engineered to mini- mize, or even avoid, untoward immune reactions. Another reason for humanizing mouse antibodies arises from the very short half life (a few hours) of mouse IgG antibodies in humans compared with the three-week half lives of their human or hu- manized counterparts. Antibody engineering has also con- tributed to the therapy of other malignan- cies such as breast cancer, which is diagnosed in more than 180,000 Ameri- can women each year. A little more than a quarter of all breast cancer patients have Some monoclonal antibodies in clinical use Monoclonal antibody [mAB] Nature of Target (Product Name) antibody (antibody specificity) Treatment for Muromonab-CD3 Mouse mAB T cells Acute rejection of liver, heart (Orthoclone OKT3) (CD3, a T cell antigen) and kidney transplants Abciximab Human-mouse Clotting receptor of platelets Blood clotting during angioplasty (ReoPro) chimeric (GP IIb/IIIa) and other cardiac procedures Daclizumab Humanized mAB Activated T cells Acute rejection of (Zenapax) (IL-2 receptor alpha subunit) kidney transplants Inflixibmab Human-mouse Tumor necrosis factor, (TNF) a Rheumatoid arthritis (Remicade) chimeric mediator of inflammation. (TNF) and Crohn’s disease Palivizumab Humanized mAB Respiratory Syncytial Virus (RSV) RSV infection in (Synagis) (F protein, a component of RSV) children, particularly infants Gemtuzumab Humanized mAB Many cells of the myeloid lineage Acute myeloid (Mylotarg) (CD33, an adhesion molecule) leukemia (AML) Alemtuzumab Humanized mAB Many types of leukocytes B cell chronic (Campath) (CD52 a cell surface antigen) lymphocytic leukemia Trastuzumab Humanized mAB An epidermal growth factor HER2 receptor-positive (Herceptin) receptor (HER2 receptor) advanced breast cancers Rituximab Humanized mAB B cells Relapsed or refractory (Rituxan) (CD20 a B cell surface antigen) non-Hodgkins lymphoma Ibritumomab Mouse mAB B cells Relapsed or refractory (Zevalin) (CD20, a B cell surface antigen) non-Hodgkins lymphoma SOURCE: Adapted from P. Carter. 2001. Improving the efficacy of antibody-based cancer therapies. Nature Reviews/Cancer 1:118. 8536d_ch05_105-136 8/23/02 11:51 AM Page 131 mac100 mac 100: 1268_tm:8536d:Goldsby et al. / Immunology 5e-: immune response. As we shall see in Chapter 11, the in vivo evolution of most humoral immune responses produces two desirable outcomes. One is class switching, in which a variety of antibody classes of the same specificity are produced. This is an important consideration because the class switching that occurs during an immune response produces antibodies that have the same specificity but different effector functions and hence, greater biological versatility. The other is the gen- eration of antibodies of higher and higher affinity as the re- sponse progresses. A central goal of Ig-gene library ap- proaches is the development of strategies to produce anti- bodies of appropriate affinity in vitro as readily as they are generated by an in vivo immune response. When the formi- dable technical obstacles to the achievement of these goals are overcome, combinatorial approaches based on phage libraries will allow the routine and widespread production of useful antibodies from any desired species without the limitations of immunization and hybridoma technology that currently complicate the production of monoclonal antibodies. Mice Have Been Engineered with Human Immunoglobulin Loci It is possible to functionally knock out, or disable, the heavy- and light-chain immunoglobulin loci in mouse embryonic stem (ES) cells. N. Lonberg and his colleagues followed this procedure and then introduced large DNA sequences (as much as 80 kb) containing human heavy- and light-chain gene segments. The DNA sequences contained constant-re- gion gene segments, J segments, many V-region segments, and, in the case of the heavy chain, D H segments. The ES cells containing these miniature human Ig gene loci (miniloci) are used to derive lines of transgenic mice that respond to anti- genic challenge by producing antigen-specific human anti- bodies (Figure 5-23). Because the human heavy- and light-chain miniloci undergo rearrangement and all the other diversity-generating processes, such as N-addition, P- addition, and even somatic hypermutation after antigenic challenge, there is an opportunity for the generation of a great deal of diversity in these mice. The presence of human heavy-chain minilocus genes for more than one isotype and their accompanying switch sites allows class switching as well. A strength of this method is that these completely hu- man antibodies are made in cells of the mouse B-cell lineage, from which antibody-secreting hybridomas are readily de- rived by cell fusion. This approach thus offers a solution to the problem of producing human monoclonal antibodies of any specificity desired. SUMMARY a73 Immunoglobulin H9260 and H9261 light chains and heavy chains are encoded by three separate multigene families, each con- taining numerous gene segments and located on different chromosomes. a73 Functional light-chain and heavy-chain genes are gener- ated by random rearrangement of the variable-region gene segments in germ-line DNA. a73 V(D)J joining is catalyzed by the recombinase activiating genes, RAG-1 and RAG-2, and the participation of other enzymes and proteins. The joining of segments is directed by recombination signal sequences (RSS), conserved DNA sequences that flank each V, D, and J gene segment. a73 Each recombination signal sequence contains a conserved heptamer sequence, a conserved nonamer sequence, and either a 12-bp (one-turn) or 23-bp (two-turn) spacer. During rearrangement, gene segments flanked by a one- turn spacer join only to segments flanked by a two-turn spacer, assuring proper V L -J L and V H -D H -J H joining. 132 PART II Generation of B-Cell and T-Cell Responses V L –C L Plasma cell #1 Promoter EcoRI V L –C L Not I Promoter EcoRINot I Prepare random combinational libraries V L –C L V H –C H 1 Plasma cell #N Promoter EcoRI V L –C L Not I EcoRI V H –C H 1 Not I Insert into λ vectors to make light- and heavy- chain libraries Amplify by PCR Isolate mRNA's Promoter Promoter EcoRINot I V L –C L Promoter EcoRINot I V L –C L Not I Not I Heavy-light construct Heavy-light construct V H –C H 1 V H –C H 1 V H –C H 1 V H –C H 1 FIGURE 5-22 General procedure for producing gene libraries en- coding Fab fragments. In this procedure, isolated mRNA that en- codes heavy and light chains is amplified by the polymerase chain reaction (PCR) and cloned in H9261 vectors. Random combinations of heavy- and light-chain genes generate an enormous number of heavy-light constructs encoding Fab fragments with different anti- genic specificity. [Adapted from W. D. Huse et al., 1989, Science 246:1275.] 8536d_ch05_105-136 8/22/02 3:07 PM Page 132 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: a73 Immunoglobulin gene rearrangements occur in sequential order, heavy-chain rearrangements first, followed by light- chain rearrangements. Allelic exclusion is a consequence of the functional rearrangement of the immunoglobulin DNA of only one parental chromosome and is necessary to assure that a mature B cell expresses immunoglobulin with a single antigenic specificity. a73 The major sources of antibody diversity, which can gener- ate H1102210 10 possible antibody combining sites, are: random joining of multiple V, J, and D germ-line gene segments; random association of heavy and light chains; junctional flexibility; P-addition; N-addition; and somatic mutation. a73 After antigenic stimulation of mature B cells, class switch- ing results in expression of different classes of antibody (IgG, IgA, and IgE) with the same antigenic specificity. a73 Differential RNA processing of the immunoglobulin heavy-chain primary transcript generates membrane- bound antibody in mature B cells, secreted antibody in Organization and Expression of Immunoglobulin Genes CHAPTER 5 133 FIGURE 5-23 Grafting human heavy- and light-chain miniloci into mice. The capacity of mice to rearrange Ig heavy- and light- chain gene segments was disabled by knocking out the C H9262 and C H9260 loci. The antibody-producing capacity of these mice was reconsti- tuted by introducing long stretches of DNA incorporating a large part of the human germ-line H9260 and heavy-chain loci (miniloci). Chimeric mice were then bred to establish a line of transgenic mice bearing both heavy- and light-chain human miniloci. Immunization of these mice results in the production of human antibody specific for the target antigen. [N. Lonberg et al., 1994, Nature 368:856.] Mouse embryonic stem cells (ES cell) Mouse ES cells incorporating human H and L miniloci Transfect into ES cells V H genes Germ-line human heavy-chain minilocus Germ-line human κ light-chain minilocus D genes J genes C μ C γ 1 V H genes J κ genes C κ μ/κ-knockout ES cells Κnockout mouse μ and κ Inject into host embryo Chimeric mouse Miniloci transgenic mouse Nontransgenic offspring Breed Blastocyst Human antibodies Immunize Human miniloci Go to www.whfreeman.com/immunology Self-Test Review and quiz of key terms 8536d_ch05_105-136 8/22/02 3:07 PM Page 133 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: plasma cells, and the simultaneous expression of IgM and IgD by mature B cells. a73 Transcription of immunoglobulin genes is regulated by three types of DNA regulatory sequences: promoters, en- hancers, and silencers. a73 Growing knowledge of the molecular biology of im- munoglobulin genes has made it possible to engineer anti- bodies for research and therapy. The approaches include chimeric antibodies, bacteriophage-based combinatorial libraries of Ig-genes, and the transplantation of extensive segments of human Ig loci into mice. References Chen, J., Y. Shinkai, F. Young, and F. W. Alt. 1994. Probing im- mune functions in RAG-deficient mice. Curr. Opin. Immunol. 6:313. Cook, G. P., and I. M. Tomlinson. 1995. The human im- munoglobulin V H repertoire. Immunol. Today 16:237. Dreyer, W. J., and J. C. Bennett. 1965. The molecular basis of an- tibody formation. Proc. Natl. Acad. Sci. U.S.A. 54:864. Fugmann, S. D., I. L. Lee, P. E. Shockett, I. J. Villey, and D. G. Schatz. 2000. The RAG proteins and V(D)J recombination: Complexes, ends and transposition. Annu.Rev.Immunol. 18:495. Gavilondo, J. V., and J. W. Larrick. 2000. Antibody engineering at the millennium. Biotechniques 29:128. Hayden, M. S., L. K. Gilliand, and J. A. Ledbetter. 1997. Antibody engineering. Curr. Opin. Immunol. 9:201. Hesslein, D. G., and D. G. Schatz. 2001. Factors and forces con- trolling V(D)J recombination. Adv. Immunol. 78:169. Hozumi, N., and S. Tonegawa. 1976. Evidence for somatic re- arrangement of immunoglobulin genes coding for variable and constant regions. Proc. Natl. Acad. Sci. U.S.A. 73:3628. Maloney, D. G., et al. 1997. IDEC-C2B8 (Rituximab) anti-CD20 monoclonal antibody therapy in patients with relapsed low- grade non-Hodgkin’s lymphoma. Blood 90:2188. Manis, J. P., M. Tian, and F. W. Alt. 2002. Mechanism and control of class-switch recombination. Trends Immunol. 23:31. Matsuda, F., K. Ishii, P. Bourvagnet, Ki Kuma, H. Hayashida, T. Miyata, and T. Honjo. 1998. The complete nucleotide sequence of the human immunoglobulin heavy chain variable region lo- cus. J. Exp. Med. 188:2151. Max, E. E. 1998. Immunoglobulins: molecular genetics. In Fun- damental Immunology, 4th ed., W. E. Paul, ed. Lippincott- Raven, Philadelphia. Mills, F. C., N. Harindranath, M. Mitchell, and E. E. Max. 1997. Enhancer complexes located downstream of both human im- munoglobulin C alpha genes. J. Exp. Med. 186:845. Oettinger, M. A., et al. 1990. RAG-1 and RAG-2, adjacent genes that synergistically activate V(D)J recombination. Science 248:1517. Tonegawa, S. 1983. Somatic generation of antibody diversity. Nature 302:575. Van Gent, D. C., et al. 1995. Initiation of V(D)J recombination in a cell-free system. Cell 81:925. Winter, G., and C. Milstein. 1990. Man-made antibodies. Nature 349:293. USEFUL WEB SITES http://www.mrc-cpe.cam.ac.uk/imt-doc/public/ INTRO.html#maps V BASE: This database and informational site is maintained at the MRC Centre for Protein Engineering in England. It is an excellent and comprehensive directory of information on the human germ-line variable region. http://www.mgen.uni-heidelberg.de/SD/SDscFvSite.html The Recombinant Antibody Page: This site has a number of links that provide interesting opportunities to explore the po- tential of genetic engineering of antibodies. http://www.ebi.ac.uk/imgt/hla/intro.html. The IMGT site contains a collection of databases of genes rel- evant to the immune system. The IMGT/LIGM database houses sequences belonging to the immunoglobulin super- family and of T cell antigen receptor sequences. Study Questions CLINICAL FOCUS QUESTION The Clinical Focus section includes a table of monoclonal antibodies approved for clinical use. Two, Rituxan and Zevalin, are used for the treatment of non- Hodgkins lymphoma. Both target CD20, a B-cell surface antigen. Zevalin is chemically modified by attachment of radioactive iso- topes (yttrium-90, a H9252 emitter or indium-111, a high energy H9253 emitter) that lethally irradiate cells to which the monoclonal an- tibody binds. Early experiments found that Zevalin without a ra- dioactive isotope attached was an ineffective therapeutic agent, whereas unlabeled Rituxan, a humanized mAB, was effective. Furthermore, Rituxan with a radioactive isotope attached was too toxic; Zevalin bearing the same isotope in equivalent amounts was far less toxic. Explain these results. (Hint: The longer a radioactive isotope stays in the body, the greater the dose of radiation absorbed by the body.) 1. Indicate whether each of the following statements is true or false. If you think a statement is false, explain why. a. V H9260 gene segments sometimes join to C H9261 gene segments. b. With the exception of a switch to IgD, immunoglobulin class switching is mediated by DNA rearrangements. c. Separate exons encode the transmembrane portion of each membrane immunoglobulin. d. Although each B cell carries two alleles encoding the im- munoglobulin heavy and light chains, only one allele is expressed. 134 PART II Generation of B-Cell and T-Cell Responses 8536d_ch05_105-136 8/22/02 3:07 PM Page 134 mac46 mac46:1256_des:8536d:Goldsby et al. / Immunology 5e: e. Primary transcripts are processed into functional mRNA by removal of introns, capping, and addition of a poly-A tail. f. The primary transcript is an RNA complement of the coding strand of the DNA and includes both introns and exons. 2. Explain why a V H segment cannot join directly with a J H seg- ment in heavy-chain gene rearrangement. 3. Considering only combinatorial joining of gene segments and association of light and heavy chains, how many differ- ent antibody molecules potentially could be generated from germ-line DNA containing 500 V L and 4 J L gene segments and 300 V H , 15 D H , and 4 J H gene segments? 4. For each incomplete statement below (a–g), select the phrase(s) that correctly completes the statement. More than one choice may be correct. a. Recombination of immunoglobulin gene segments serves to (1) promote Ig diversification (2) assemble a complete Ig coding sequence (3) allow changes in coding information during B-cell maturation (4) increase the affinity of immunoglobulin for antibody (5) all of the above b. Somatic mutation of immunoglobulin genes accounts for (1) allelic exclusion (2) class switching from IgM to IgG (3) affinity maturation (4) all of the above (5) none of the above c. The frequency of somatic mutation in Ig genes is greatest during (1) differentiation of pre-B cells into mature B cells (2) differentiation of pre-T cells into mature T cells (3) generation of memory B cells (4) antibody secretion by plasma cells (5) none of the above d. Kappa and lambda light-chain genes (1) are located on the same chromosome (2) associate with only one type of heavy chain (3) can be expressed by the same B cell (4) all of the above (5) none of the above e. Generation of combinatorial diversity among im- munoglobulins involves (1) mRNA splicing (2) DNA rearrangement (3) recombination signal sequences (4) one-turn/two-turn joining rule (5) switch sites f. A B cell becomes immunocompetent (1) following productive rearrangement of variable- region heavy-chain gene segments in germ-line DNA (2) following productive rearrangement of variable- region heavy-chain and light-chain gene segments in germ-line DNA (3) following class switching (4) during affinity maturation (5) following binding of T H cytokines to their receptors on the B cell g. The mechanism that permits immunoglobulins to be synthesized in either a membrane-bound or secreted form is (1) allelic exclusion (2) codominant expression (3) class switching (4) the one-turn/two-turn joining rule (5) differential RNA processing 5. What mechanisms generate the three hypervariable regions (complementarity-determining regions) of immunoglobu- lin heavy and light chains? Why is the third hypervariable re- gion (CDR3) more variable than the other two (CDR1 and CDR2)? 6. You have been given a cloned myeloma cell line that secretes IgG with the molecular formula H9253 2 H9261 2 . Both the heavy and light chains in this cell line are encoded by genes derived from allele 1. Indicate the form(s) in which each of the genes listed below would occur in this cell line using the following symbols: G H11005 germ line form; R H11005 productively rearranged form; NP H11005nonproductively rearranged form. State the rea- son for your choice in each case. a. Heavy-chain allele 1 d. H9260-chain allele 2 b. Heavy-chain allele 2 e. H9261-chain allele 1 c. H9260-chain allele 1 f. H9261-chain allele 2 7. You have a B-cell lymphoma that has made nonproductive rearrangements for both heavy-chain alleles. What is the arrangement of its light-chain DNA? Why? 8. Indicate whether each of the class switches indicated below can occur (Yes) or cannot occur (No). a. IgM to IgD d. IgA to IgG b. IgM to IgA e. IgM to IgG c. IgE to IgG 9. Describe one advantage and one disadvantage of N- nucleotide addition during the rearrangement of im- munoglobulin heavy-chain gene segments. 10. X-ray crystallographic analyses of many antibody molecules bound to their respective antigens have revealed that the CDR3 of both the heavy and light chains make contact with the epitope. Moreover, sequence analyses reveal that the variability of CDR3 is greater than that of either CDR1 or CDR2. What mechanisms account for the greater diversity in CDR3? 11. How many chances does a developing B cell have to generate a functional immunoglobulin light-chain gene? 12. Match the terms below (a–h) to the description(s) that fol- low (1–11). Each description may be used once, more than once, or not at all; more than one description may apply to some terms. Organization and Expression of Immunoglobulin Genes CHAPTER 5 135 8536d_ch05_105-136 8/1/02 8:53 AM Page 135 mac79 Mac 79:45_BW:Goldsby et al. / Immunology 5e: Terms a. ______ RAG-1 and RAG-2 e. ______ P-nucleotides b. ______ Double-strand break f. ______ N-nucleotides repair (DSBR) enzymes g. ______ Promoters c. ______ Coding joints h. ______ Enhancers d. ______ RSSs Descriptions (1) Junctions between immunoglobulin gene segments formed during rearrangement (2) Source of diversity in antibody heavy chains (3) DNA regulatory sequences (4) Conserved DNA sequences, located adjacent to V, D, and J segments, that help direct gene rearrangement (5) Enzymes expressed in developing B cells (6) Enzymes expressed in mature B cells (7) Nucleotide sequences located close to each leader seg- ment in immunoglobulin genes to which RNA poly- merase binds (8) Product of endonuclease cleavage of hairpin interme- diates in Ig-gene rearrangement (9) Enzymes that are defective in SCID mice (10) Nucleotide sequences that greatly increase the rate of transcription of rearranged immunoglobulin genes compared with germ-line DNA (11) Nucleotides added by TdT enzyme 13. Many B-cell lymphomas express surface immunoglobulin on their plasma membranes. It is possible to isolate this lym- phoma antibody and make a high affinity, highly specific mouse monoclonal anti-idiotype antibody against it. What steps should be taken to make this mouse monoclonal anti- body most suitable for use in the patient. Is it highly likely that, once made, such an engineered antibody will be gener- ally useful for lymphoma patients? 136 PART II Generation of B-Cell and T-Cell Responses 8536d_ch05_105-136 8/1/02 8:53 AM Page 136 mac79 Mac 79:45_BW:Goldsby et al. / Immunology 5e: