The human heavy chain diseases (HCDs) are proliferative disorders of B lymphoid cells that produce truncated monoclonal immunoglobulin heavy chains lacking associated light chains [1,2]. μ, γ and α HCDs have been described. Structural studies have shown that most HCD proteins have internal deletions [1,3] beginning at various points within the variable region and extending through the first domain (CH1) of the constant region. Normal sequence is usually resumed at the beginning of the hinge or at well defined interdomain regions [1-4]. This feature is of particular interest because of the discontinuous organization of the genomic IGHC genes where each constant region domain is encoded by a separate exon [5-8]. In addition to the defective heavy chains, most cells producing γ and α HCD proteins also fail to synthesize light chains. Biosynthetic studies have discounted extensive post-synthetic degradation to explain the truncated HCD proteins . Thus the abnormal heavy chains must be the result of either aberrant gene rearrangements or of mutational events leading to altered RNA splicing or incorrect transcription .
Structural analysis of the μ HCD present in the serum of the patient BW had revealed the presence of a monomeric heavy chain with a molecular weight of 58 000 which lacked associated light chains . The entire variable region was deleted with the N-terminal sequence of the constant region being initiated at the fifth amino acid of the first domain of the IGHM (Cμ) region (Figure 1). Post-synthetic degradation was not occurring because the cytoplamic and the secreted μ chains, as well as the primary in vitro translation product were similarly truncated . The molecular defect was therefore at the DNA level.
Cloning and sequencing showed a IGHV-IGHD-IGHJ4 rearranged gene (Figure 1). A small DNA insertion/deletion eliminated the J4 donor splice site at the 3 end of the VDJ rearrangement and necessitated an alternate RNA splicing site between the leader and the CH1 exon . This molecular defect generates truncated μ chains present both as membrane and as secreted molecules that lack the variable region and fail to associate with light chain.
The γ3 HCD protein isolated from the serum of patient OMM revealed a monomeric chain with a molecular weight of 40 000 and no associated light chains. The N-terminal sequence corresponded to the beginning of the normal IGHG3 hinge . However examination of the cDNA sequence shows that it encodes the first 15 residues of the V region and that it is compatible with an extensive internal deletion encompassing the remainder of the V and the entire CH1 domain . Another possibility might be a mutation or a small deletion resulting in an altered or missing normal splice site and responsible for an alternate splice which would join two coding regions that are not normally contiguous.
A series of mouse myeloma mutants derived from a cell-line of the murine MPC-11 tumor (γ2b, kappa) resemble human HCDs in their loss of an internal domain. In these mutants (for example, 10.1 in Figure 1), most of the IGHG2B (γ2b) CH1 exon was present in the nuclear RNA but was removed during splicing to form the mature cytoplasmic RNA. Sequence studies showed a deletion of 99 nucleotides at the 3 end of the CH1 exon which removed the donor splice site, apparently causing the aberrant splicing of the RNA transcript .
A shortened γ1-chain lacking also CH1 sequences is produced by the murine 1F2 (a mutant cell line of the murine MOPC21 tumor). Cloning and sequencing of the corresponding IGHG1 gene have shown that it contains an extensive deletion which removes the entire CH1 exon and a large portion of the intervening sequences (IVS) between J and CH1 (Figure 1). Due to this deletion the switch (S) sequence is adjacent to the hinge segment and, consequently, an aberrant switch recombination event  is responsible for the direct splicing of the VDJ exon to the hinge exon. An interesting feature of this deleted IGHG1 gene is its strong resemblance with the human IGHEP1 gene .
These two mouse mutant (10.1 and 1F2) examples show that different deletions at the DNA level result in a similar aberrant splicing and in the synthesis of mRNAs and proteins displaying similar internal deletions.
We will now describe an example where an aberrant rearrangement leads to the abnormal splicing. In the mouse MPC-11 tumor cell line, one IGKV gene recombines to a heptanucleotide signal between IGKJ5 and IGKC  (Figure 1). This aberrant rearrangement removes the five IGKJ genes (and therefore their donor splice sites). When the precursor transcript is processed, the L-PART1 is directly spliced to the IGKC gene and the entire V region is deleted resulting in a truncated kappa chain [15-17]. This abnormal kappa chain is similar to the human μ HCD from BW  which also lacks the variable region.
Mouse α HCDs have been described in two mutants of the murine myeloma W3129. These mutants produce α chain protein with similar internal deletions and no light chain. Both mutants result from small genomic DNA deletions of different extents which are felt to arise by a recombination/excision mechanism between region of homology within the IGHA gene . Although these deletions are of different extents (one encompassing the entire CH1 exon, the other only part of it), their transcripts are processed to give rise to shortened mRNA and proteins of the same sizes. Internal deletions of CH exons have also been noted in the characterization of shortened μ chains of mutant mouse hybridomas .
Molecular analysis of HCDs or mouse variants exclude the post-synthetic degradation (that is the degradation of an initially normal immunoglobulin chain) as the mechanism responsible for the truncated protein. In most cases, abnormal RNA splicing provides an explanation for the relative constancy with which normal sequence is resumed at an exon boundary in HCD proteins. This aberrant RNA splicing results from DNA defects, such as small or large genomic deletions (like in 10.1), mutations, aberrant switch event (like in 1F2) or aberrant rearrangement (like in MPC-11 K), which eliminate crucial donor or acceptor sites.
In the HCDs proteins, abnormal RNA splicing allows to bypass an eliminated splicing site, provided that the RNA precursor contains other introns with functional splicing sites. Such events result in truncated proteins which are missing exon information. This seems to be the most likely explanation for the large number of IG heavy chain deletion variants which have lost internal domains or the hinge exon .
Most γ and α HCDs fail to produce light chain while μ HCDs synthetize free but unassociated light chains. The defects responsible for the absence of light chains in most γ and α HCDs or the lack of heavy light chain assembly in μ HCDs are not readily explained. The heavy chain defects may have a secondary effect upon light chain genes since in a murine example, the heavy chain disease arose in two steps : first a deleted heavy and a normal light chain synthesis or assembly. Alternatively, a second independent genetic defect may occur in the light chain gene since in a murine example, the heavy chain disease arose in two steps : first a deleted heavy and a normal light chain were synthesized and then a secondary variant synthesizing only the deleted heavy chain was identified.
In most HCDs, the absence of the heavy chain CH1 domain may be responsible for the lack of assembly of the light chain. This is remarkably similar to the situation found for the camel gamma heavy chains (IMGT Repertoire).