Polymerization of secretory IgM in B lymphocytes is prevented by a prior targeting to a degradation pathway.

B lymphocytes express on their surface a membrane form of IgM (mIgM), and synthesize but fail to secrete a secretory form of IgM (sIgM). Plasma cells shift to the exclusive synthesis and efficient secretion of sIgM. The sIgM in B cells differs from that in plasma cells in its pattern of assembly: in plasma cells, monomers of sIgM are assembled into polymers and only polymers are secreted; in B lymphocytes, monomeric sIgM is neither polymerized nor secreted and is degraded intracellularly. In this article we blocked the export of proteins from the endoplasmic reticulum at low temperatures or with energy poisons or brefeldin A, and localized the different assembly steps of mIgM and sIgM in the 38C B lymphocytes and of sIgM in the 38C-derived sIgM-secreting D2 hybridoma. In both cell lines, sIgM assembly into monomers was not affected, whereas polymerization of sIgM in D2 cells and monomer formation of mIgM in 38C cells were strongly inhibited. Moreover, probing with specific lectins revealed galactosylated monomers and polymers in D2 cells and galactosylated hemimer and monomers only of mIgM in 38C cells. In addition, when Golgi functions were hampered with Tris base, monomerization of mIgM and polymerization of sIgM were attenuated. These results indicate that polymerization of sIgM in D2 cells and monomerization of mIgM in 38C cells are post-endoplasmic reticulum events, occurring in or beyond the trans-Golgi galactosylation compartment. Since only polymers are secreted from D2 cells and only monomeric mIgM is displayed on the surface of 38C cells, partially assembled molecules may traverse the secretory pathway yet are restricted from the cell surface. Furthermore, monomeric sIgM in 38C cells is never galactosylated, thus it is degraded prior to the galactosylation compartment. We conclude that targeting of sIgM to degradation in 38C cells precedes its assembly site into polymers in D2 cells. This implies that degradation of sIgM does not result from the incompetence of 38C cells to polymerize. Rather, assembly of sIgM into polymers and their subsequent secretion are prevented in B lymphocytes by preceding targeting of monomeric sIgM to degradation.     

Degradation of secretory immunoglobulin M in B lymphocytes occurs in a postendoplasmic reticulum compartment and is mediated by a cysteine protease.

In 38C B lymphocytes, membrane IgM is expressed on the surface, whereas secretory IgM (sIgM) is rapidly degraded. Here, we localize this degradation and characterize the proteases involved in this process. Upon treatment with brefeldin A, degradation of sIgM in 38C cells was strongly inhibited, as was secretion from the sIgM-secreting D2 hybridoma. Moreover, the brefeldin A-induced Golgi resorption resulted in galactosylation of sIgM and partial resistance to endoglycosidase H. However, sIgM avoided degradation neither due to modified terminal glycosylation nor as a consequence of the brefeldin A-induced altered milieu of the endoplasmic reticulum. When these modifications were prevented by inhibiting retrograde transport with nocodazole or by abrogating terminal glycosylation with swainsonine, sIgM was still rescued from degradation. The unaffected breakdown in the presence of nocodazole also argued against recycling of sIgM to be degraded in the endoplasmic reticulum. Furthermore, upon removal of brefeldin A, degradation of galactosylated sIgM resumed in 38C cells, as did secretion from D2 cells. These results indicate that functional export of proteins from the endoplasmic reticulum is a prerequisite for sIgM degradation. Biochemical characterization of this novel postendoplasmic reticulum/pre-trans-Golgi proteolytic pathway included application of inhibitors to a broad spectrum of proteases. Among the compounds tested, only calpain inhibitor I exerted strong inhibition. The involvement of cysteine protease(s) in the degradation of sIgM was corroborated by the inhibitory effect of diamide. We conclude that B lymphocytes avoid secretion by active and selective targeting of sIgM to a developmentally regulated postendoplasmic reticulum degradation pathway in which degradation is mediated by a cysteine protease.     

Expression, distribution and specificity of Fc receptors for IgM on murine B cells

Subpopulations of normal adult murine splenic B cells and a panel of murine B cell tumors were examined for their ability to bind murine IgM specifically. By using two-color flow cytometric analyses, we have demonstrated that 90 to 95% of surface (s)IgD+ B cells express surface membrane receptors for IgM (Fc mu R). The binding of pentameric murine IgM to splenocyte Fc mu R was IgM-specific since it was totally inhibited by other polymeric IgM proteins, but not by Ig of other H chain classes or by mAb specific for the murine IgG or IgE FcR. Binding of IgM to splenic cells was saturable. Fc mu R were co-expressed with the Fc gamma R as well as the Fc epsilon R on the majority of splenic B cells. Minor populations of splenic mononuclear cells expressed only an Fc mu R, Fc gamma R or Fc epsilon R. In a survey of B tumor cell lines representing different stages of B cell development, we observed that the Fc mu R was expressed on pre-B cell lines and that Fc mu R detection was maximal on immature B cell lines that expressed sIgM and low amounts of sIgD and Ia. Fc mu R were not detected on cell lines that had switched from sIgM to the expression of another sIg, or on plasmacytomas and hybridomas. The studies with normal splenocytes establish that the majority of sIgD+ B lymphocytes in adult BALB/c mice express surface membrane receptors that specifically bind IgM. The studies with B lineage tumor cells suggest that the expression of Fc mu R on B cells is developmentally regulated and that the pattern of expression exhibited by Fc mu R during B cell ontogeny differs from the patterns that have been previously found for IgG and IgE FcR. These observations raise the possibility that Fc mu R might have a functional significance in some aspect of B cell maturation and activation. By using a family of IgM H chain constant region domain deletional mutants, we have further demonstrated that, like the T cell Fc mu R, the B cell Fc mu R also requires a C mu 3 domain for binding to occur, raising the possibility that the T and B cell Fc mu R in mice may be structurally related molecules.     

Control of IgM synthesis in the murine pre-B cell line, 70Z/3'

The murine 70Z/3 tumor resembles a pre-B cell in synthesizing only intracellular mu-chains and no detectable light chain. However, one kappa gene is already rearranged, and after overnight incubation with lipopolysaccharide (LPS), most of the cells are induced to synthesize light chain. The induced cells display IgM on their surface, but do not secrete IgM. Thus, 70Z/3 cells resemble cells poised at the pre-B cell/B lymphocyte border. We have examined synthesis and post-translational modification of mu-chains in uninduced and induced 70Z/3 cells. Isolation of mu-chains and peptide maps demonstrated that both populations synthesize intracellular forms that correspond to membrane-specific mum and secretion-specific mus chains. These intracellular forms have completed only the first of the two glycosylation steps characteristic of eukaryotic cells. After induction by LPS, L chain synthesis commences, mum and mus synthesis are both increased twofold to threefold (due to an increased rate of synthesis rather than decreased degradation), and both complex with L chain to form mu2L2 tetramers. Furthermore, the glycosylation of a subset of the mum chains is completed, and these are placed on the membrane. However, unglycosylated mu2L2 tetramers can be placed on the membrane, so glycosylation is not a requirement. These data suggest that L chain may not be sufficient for externalization of mum and mus chains. These data support the idea that the controls of membrane placement and secretion of mu chains are post-translational and that different mechanisms operate for mum and mus chains.     

Mouse pre-B cells synthesize and secrete mu heavy chains but not light chains

The immunoglobulins produced by the earliest recognizable B cell precursors (pre-B cells) were characterized in the mouse and human. Immunofluorescent analysis revealed no evidence of surface IgM components, and only mu heavy chains could be detected intracytoplasmically in pre-B cells. Surface IgM components could not be isolated from intact fetal liver cells that lacked sIgM+ B lymphocytes but possessed pre-B cells. Pre-B cells were shown to synthesize and secrete mu heavy chains but not light chains by immunochemical analysis. These mu chains constituted less than 0.01% of TCA precipitable protein synthesized and secreted by fetal liver cells during an 8 hr labelling period. Migration of both intracellular and secreted mu chains on SDS-PAGE suggested that they were smaller than mu chains secreted by mouse and human plasmacytomas. These data indicate that mu chain synthesis precedes light chain expression during B cell ontogeny and suggest a new role for pre-B cells in the generation and expression of a diverse immunoglobulin repertoire.     

Aberrant trafficking of the B cell receptor Ig-alpha beta subunit in a B lymphoma cell line

The B cell Ag receptor (BCR) has two important functions: first, it binds and takes up Ag for presentation to T lymphocytes; and second, it transmits signals that regulate B cell development. Normal expression of the BCR requires the association of the Ag binding subunit, membrane IgM (mIgM), with the signaling component, the Ig-alpha beta heterodimer. After assembly in the endoplasmic reticulum, the intact BCR travels through the secretory pathway to the cell surface. In this paper, we report two variants of the B lymphoma cell lines, WEHI 279 and WEHI 231, that have both lost the ability to express mu heavy chain and consequently do not express mIgM. However, these variants do express the Ig-alpha beta heterodimer. In one variant, WEHI 279*, the Ig-alpha beta remained trapped intracellularly in the absence of mIgM. The other variant, 303.1.5.LM, expressed an aberrantly glycosylated Ig-alpha beta on the cell surface that was capable of signaling after cross-linking with anti-Ig-beta Abs. Further characterization uncovered a point mutation in the 303.1.5.LM mb1 gene that would change a proline for a leucine in the extracellular domain of Ig-alpha. The 303.1.5.LM Ig-alpha beta could not associate with a wild-type mIgM after mu heavy chain was reconstituted by DNA transfection. Thus, this mutation could define a region of the Ig-alpha polypeptide that is important for recognition by the endoplasmic reticulum quality control system, for association with glycosylating enzymes, and for the association of Ig-alpha beta subunits with mIgM subunits to create a complete BCR complex.     

Polyclonal activation of the murine immune system by an antibody to IgD. VIII. Stimulation of IgD secretion

The low levels of serum IgD found in mice and the lack of a typical DNA switch sequence between C delta and C mu raise the possibility that the generation of murine IgD-secreting cells results from a chance "mistake" rather than a controlled process. The recent observation that injection of mice with purified IgD upregulates IgD receptor expression on helper T cells and enhances the ability of these T cells to induce B cells to differentiate into antibody secreting cells led us to look for evidence of controlled differentiation of B cells into IgD-secreting cells. To do this, we injected mice with a goat antibody to IgD (GaM delta), because this antibody stimulates large increases in IgM, IgG1, IgG2a, and IgE secretion. Mice injected with GaM delta demonstrated a large increase in splenic content of mRNA specific for the secreted form of delta-chain, as well as a greater than 100-fold increase in the percentage of splenic IgD-containing plasmablasts. The secretory IgD response was totally T-dependent. Production of the secretory form of IgD was not seen until 7 days after GaM delta injection, and peaked sharply on day 8, whereas by day 6 IgM secretion had already peaked and IgG1 and IgG2 secretion had attained substantial levels. This observation suggests that: 1) either cells that synthesize large quantities of the secretory form of delta-chain, unlike cells that synthesize large quantities of the secretory forms of gamma-, epsilon-, or alpha-chains, do this without deleting C mu, or, despite the absence of a typical DNA switch sequence between C mu and C delta, controls must exist to effect the C mu deletion and VDJ-C delta joining; and 2) if secreted IgD has a role in the regulation of a humoral immune response it most likely is involved in later processes, such as memory cell generation or response termination, rather than in relatively early processes, such as helper T cell activation.     

Regulation of membrane IgM expression in secretory B cells: translational and post-translational events.

IgM secreting cells express little or no membrane IgM. This is not always due to absence of the relevant mRNA. To investigate the synthesis and processing of membrane (micron) and secreted (microseconds) polypeptides in secretory B cells, myeloma cells were transfected either with a plasmid containing an intact mu gene or with one only capable of directing micron (not microseconds) mRNA synthesis. Although myeloma transfectants could make abundant levels of micron mRNA, they did not express IgM on the cell surface. In the myeloma host, micron mRNA is translated some 5-fold less efficiently than microseconds mRNA. However, this translational control does not totally preclude micron synthesis, indicating post-translational regulatory events. No difference between micron and microseconds chains could be detected in their rate of assembly with light chains or in their stability, although both types of heavy chain were degraded more rapidly when synthesized in the absence of light chain, or when the hydrophobic nature of the leader sequence was destroyed by site-directed mutagenesis. However, whereas intracellular microseconds chains in IgM-secreting plasmacytoma were found to be concentrated in the Golgi, the micron chains were mainly located in the endoplasmic reticulum. Retention in the endoplasmic reticulum is also observed for both micron and microseconds when synthesized in the absence of light chain. We propose that it is the expansion of the endoplasmic reticulum that accompanies B cell to plasma cell differentiation which is in part responsible for the down-regulation of surface IgM expression. Such a mechanism may also affect the expression of other surface proteins.     

Transient down-regulation of PKC-zeta RNA following crosslinking of membrane IgM on WEHI-231 B lymphoma cells.

Regulation of protein kinase C (PKC) isoform mRNAs has been studied in the immature, murine B lymphoma WEHI-231 by the MAPPing protocol and by slot blot analysis of unamplified mRNA. This membrane IgM (mIgM)-positive cell line has been previously used as a model to study signal transduction by mIgM in immature B lymphocytes and the role of those signals in the induction of immune tolerance in the B cell compartment. Stimulation of the cells by anti-mu antibodies, phorbol ester, or Ca2+ ionophore caused growth arrest and death of the cells. IL 4 and IL 5 slowed the growth of the cells. Of these stimuli, only anti-mu stimulation affected PKC mRNA levels. Anti-mu treatment caused a transient decrease in the amount of PKC-zeta isoform mRNA within 3 hr. Within 24 hr levels returned toward normal. Anti-mu had little or no effect on the expression of mRNA for the alpha, beta, delta, or epsilon isoforms of PKC. WEHI-231 cells do not express PKC-gamma. Although anti-mu treatment blocked progression of the cells from the G0/G1 stage into the S phase of cell cycle, viable sort selected cells in either the G0/G1 or the S/G2/M phases showed no clear difference in the expression of PKC-zeta message. Thus, there is not preferential regulation of expression of PKC-zeta during stages of the cell cycle. The results show that mIgM on WEHI-231 cells can transduce a signal that is not mediated by PKC or Ca2+ mobilization alone. The signal causes transient, selective down-regulation of mRNA encoding the zeta PKC isoform.     

Immune dysfunction in mice with plasmacytomas. I. Evidence that transforming growth factor-beta contributes to the altered expression of activation receptors on host B lymphocytes.

Plasmacytoma-bearing mice (PC-mice) develop a polyclonal B cell immunodeficiency syndrome characterized by marked impairment of: a) primary antibody responses and b) proliferative responses to B cell mitogens. The present investigations used two-color flow cytometry to examine B lymphocytes from the spleens and lymph nodes of PC-mice and found decreased surface membrane expression of surface IgM (sIgM), transferrin receptors (TfR) and IgE FcR (CD23), increased expression of class II MHC, but normal expression of B220, Mel-14, Fc gamma RII, and Fc mu R. These changes were not related to the H chain class or the amount of Ig produced by the plasmacytoma. When cultured with IL-4, B lymphocytes from PC-mice increased their expression of sIgM and class II MHC, but not of CD23. Several findings implicate transforming growth factor-beta 1 (TGF-beta 1) in the mechanism that modulates receptor expression on B lymphocytes in PC-mice: a) ascites fluid from PC-mice contains large quantities of TGF-beta 1; b) supernatants of cultured spleen cells from PC mice contain up to eightfold more TGF-beta than is found with normal spleen cells; c) cloned plasmacytoma cells produce TGF-beta in vitro; and d) the abnormal phenotype of B cells from PC-mice, i.e., decreased CD23, sIgM, and TfR, and increased class II MHC, is induced on normal B cells cultured in the presence of TGF-beta 1. Because sIgM, TfR, class II MHC, and CD23 are molecules that play fundamental roles in the activation of normal B cells, their modulation by TGF-beta 1: a) identifies molecular mechanisms that could account for some of the known immunosuppressive properties of TGF-beta 1 and b) implicates TGF-beta in the pathogenesis of the polyclonal B cell immunodeficiency that is characteristic of plasma cell tumors.     

Generation of phosphatidic acid and diacylglycerols following ligation of surface immunoglobulin in human B lymphocytes: potential role in PKC activation.

We have examined signal transduction via membrane IgM (mIgM) in resting and cycling human B cells. Crosslinking mIgM on all of the cell types studied transduced a signal through the phosphatidylinositol pathway, producing inositol 1,4,5-trisphosphate and release of intracellular free calcium. These second messengers were formed regardless of quantitative or qualitative differences in the surface expression of mIgM: cells that had low levels of surface IgM (T-51) or had no light chain associated with surface heavy chain (DB) signaled phosphatidylinositol pathway activation after mIgM crosslinking. Production of specific lipid products in nonquiescent B cells differed from that in normal resting cells. Ligation of surface immunoglobulin on resting B cells resulted in sustained increases of both diacylglycerol and phosphatidic acid, two lipids that can influence PKC activation. Whereas PKC was strongly activated in normal tonsillar B cells, several cell lines had reduced PKC activation following crosslinking of mIgM. The reduction in protein kinase C activation correlated with the absence or reduced levels of phosphatidic acid or diacylglycerol following stimulation: protein kinase C translocated and was activated only in cells that had elevated levels of both diacylglycerides and phosphatidic acid. Anti-IgM-induced phosphorylation of a protein kinase C substrate protein CD20, also increased in those cells having PKC activation and not in cells in which kinase activity was reduced. CD20 phosphorylation also increased following the direct addition of exogenous phosphatidic acid to resting B cells. Together, these observations show that the generation of lipid products following mIgM crosslinking in resting cells can vary from that in cycling cells and may relate to the different levels of PKC activation. In a companion study we report that ligation of surface IgM activates both an acyltransferase and phospholipase D to form phosphatidic acid.     

Regulation of nuclear poly(A) addition controls the expression of immunoglobulin M secretory mRNA.

B-cell differentiation is accompanied by a dramatic increase in cytoplasmic accumulation and stability of the IgM heavy chain (mu) secretory mRNA. Despite considerable effort, the mechanism is unknown. We have identified three short motifs upstream of the secretory poly(A) site, which, when mutated in the mu heavy chain gene, significantly increase the accumulation of the secretory form of poly(A)(+) mRNA relative to the membrane form and regulate the expression of the secretory poly(A) site in a developmental manner. We show that these motifs bind U1A and inhibit polyadenylation in vitro and in vivo. Overexpression of U1A in vivo results in the selective inhibition of the secretory form. Thus, this novel mechanism selectively controls post-cleavage expression of the mu secretory mRNA. We present evidence that this mechanism is used to regulate alternative expression of other genes.     

Circulating monoclonal IgM proteins in B cell chronic lymphocytic leukemia: their identification, characterization and relationship to membrane IgM

Accumulated evidence indicates that there is a circulating monoclonal Ig protein related to the leukemic cell-associated Ig in the majority of patients with B cell chronic lymphocytic leukemia (CLL) despite the failure to demonstrate such a protein by conventional serum electrophoresis. Methodology has been developed to reveal these hidden monoclonal bands and to show that they are related to the leukemia-associated membrane Ig (mIg). Of nine CLL cases with stainable mIgM and without discernable plasma Ig bands, marked hypogammaglobulinemia was evident in six. In the other three, a significant amount of protein was present in the gamma region. IgM was isolated from the plasma of these patients by affinity chromatography with Sepharose-4B, conjugated with affinity purified anti-human IgM antibodies. One to 3 mg were isolated from 20 to 40 ml of plasma. Agarose electrophoresis revealed a monoclonal Ig band in the isolated IgM in all cases. Eight of these IgM proteins were analyzed by high-pressure liquid chromatography. Five were found to be pentameric IgM. In the remaining three, various amounts of monomeric IgM were detected. Attempts to make anti-idiotypic antibodies to the isolated proteins have been successful. Thus far, a rabbit anti-idiotypic antiserum was obtained in one case and two mouse monoclonal anti-idiotypic antibodies in two additional cases. Immunofluorescence analysis revealed that plasma IgM and mIgM shared similar idiotypic determinants. One other monoclonal antibody was shown to be specific for a V region marker of a minor Ig population. These findings indicate that B leukemic lymphocytes do secrete a small amount of IgM and lend further support to the thesis that the maturation defect in CLL is incomplete. It is also feasible to isolate the secreted IgM and to produce anti-idiotypic antibodies to them. In view of the potential therapeutic effect of anti-idiotypic antibodies, this may offer an alternative and efficient approach to generate a large panel of anti-idiotypic antibodies for clinical trials. The possibility also exists that this approach is applicable to other B cell proliferative disorders such as the non-Hodgkin B cell lymphomas.     

Lactoperoxidase-catalyzed iodination of human IgM. Differences between 7 S IgM and 19 S IgM and between cell surface 7 S IgM and serum 7 S IgM.

We have studied the lactoperoxidase-catalyzed iodination of human IgM and have measured the ratio of radioactivity incorporated into the mu chain to that incorporated into the L chain (i.e. the mu/L ratio). Both 7 S and 19 S IgM were examined. The ratio of radioactivity was found to be larger for 7 S IgM than for 19 S IgM for all four of the monoclonal IgM proteins examined. The data suggest that some tyrosines of the mu chain which are buried and not available for iodination in 19 S IgM become exposed on conversion of 19 S IgM to 7 S IgM. The mu/L ratio for the IgM found on the cell surface of RPMI 8392 cells was significantly smaller than the ratios for all of the five 7 S IgM proteins studied in solution. It appears, therefore, that a portion of the mu chain of the cell surface IgM of the RPMI 8392 cells is buried in the membrane.     

Surface mu heavy chain expressed on pre-B lymphomas transduces Ca2+ signals but fails to cause growth arrest of pre-B lymphomas.

Little is known about the role of signals transduced by cell surface IgM (sIgM) expressed during early B cell development. A subclone (1.6) of the late pre-B cell lymphoma 70Z/3.12 was used to study signal transduction by surface mu heavy (H) chain before and after transition to the early immature B cell stage, and the functional consequences thereof. Although kappa L chain expression can be induced on 1.6 cells by LPS or cytokines, immunoprecipitations indicated that the non-induced 1.6 cells expressed mu H chain with an alternative protein(s) which may be a surrogate light chain(s). Consistent with this, anti-mu but not anti-kappa or anti-lambda antibodies caused transient Ca2+ mobilization in noninduced 1.6 cells. The Ca2+ signal was derived from both intracellular stores and Ca2+ influx in either noninduced cells or in cells that had been preinduced to express kappa L chain. Thus, the ability of mu H chain to mobilize Ca2+ as a second messenger does not depend upon the expression of mature L chains. The immature B lymphomas, WEHI-231 and CH1, express mature forms of IgM and undergo growth arrest when stimulated by anti-mu antibody. In contrast, signals generated by mu H chain on either noninduced or preinduced 1.6 cells or in the sIgM+ pre-B cell transfectant 300-19 mu lambda 36/8 did not cause growth arrest. These results suggest that mu H chain expressed on pre-B cells is capable of mobilizing Ca2+, but that this signal alone is insufficient to induce growth arrest in the pre-B cell.     

Immunoglobulin light chains dictate vesicular transport-dependent and -independent routes for IgM degradation by the ubiquitin-proteasome pathway

Degradation of IgM mu heavy chains in light chain-negative pre-B cells is independent of vesicular transport, as is evident by its insensitivity to brefeldin A or cell permeabilization. Conversely, by the same criteria, degradation of the secretory mu heavy chain in light chain-expressing B cells depends on vesicular transport. To investigate whether the presence of conventional light chains or the developmental stage of the B-lymphocytes dictates the degradative route taken by mu, we express in 70Z/3 pre-B cells either lambda ectopically or kappa by lipopolysaccharides-stimulated differentiation into B cells and show their assembly with mu heavy chains. The resulting sensitivity of mu degradation to brefeldin A and cell permeabilization demonstrates that conventional light chains, a hallmark of B cell differentiation, are necessary and sufficient to divert mu from a vesicular transport-independent to a vesicular transport-dependent degradative route. Although both routes converge at the ubiquitin-proteasome degradation pathway, only in light chain-expressing cells is vesicular transport a prerequisite for mu ubiquitination.     

Ontogeny of B cells expressing IgM in embryonic and larval tissues of the American grass frog, Rana pipiens

Affinity-purified, fluorochrome-tagged F(ab')(2) antibody fragments specific for heavy (mu) chains of Rana pipiens IgM were prepared from hyperimmune rabbit sera. By using two-color immunofluorescent procedures we observed that (1) the first cells expressing IgM, termed pre-B cells, lack detectable quantities of membrane or surface IgM but contain detectable quantities of cytoplasmic IgM (smu(-)/cmu(+)), (2) sIgM(+) B cells were the second type of IgM containing cell to appear in development, and (3) plasma cells, which contain copious quantities of cIgM, were the final phenotype to appear in the development of B cells expressing IgM. These cells were first observed in the pronephros of the developing urogenital system. Shortly after their appearance in the pronephros, cells in B lineages were observed in the liver. These observations (1) are consistent with recent studies of B lymphopoiesis in the aorta-gonad-mesonephros (AGM) region in endothermic vertebrates, including mice, (2) suggest that there are fundamental ontogenetic and phylogenetic similarities between cells and tissues of developing vertebrate immune systems, and (3) evoke questions concerning the possible function(s) of lymphocytes in developing anurans up to metamorphosis and beyond.