Synthesis of Multiple Immunoglobulin Classes by Single Lymphocytes

Transcription

1 Synthesis of Multiple Immunoglobulin Classes by Single Lymphocytes B. PERNIS, L. FORNI AND A. L. LUZZATI Basel Institute for Immunology, CH 4058 Basel, Switzerland There is good evidence that, in agreement with the principles of clonal selection, the immunoglobulin molecules synthesized by one B immunocyte, or by a clone of these cells, are uniform with regard to the Fv portion; that is, they have a uniform antibedy-combining site and a uniform idiotype (reviewed by Makela and Cross 1970). This condition applies beth to the membrane-bound immunoglobulins (Raft et al. 1973) and to the secreted molecules (Miikel~i 1968). Thus there are predictable restrictions to the synthesis of immunoglobul-in chains of one allotype only (allelic exclusion) since the two homologous chromosomes are not expected to carry identical sets of variable-region genes, neither germ-line genes nor genes generated by a process of somatic mutation. The same reasoning applies for normal B immunocytes, which are restricted to the synthesis of only one type of light (L) chain (K or ~,) because the two types of L chains, coded by genes on different chromosomes, do not have identical sets of variableregion genes (see Pernis 1968). On the other hand, the synthesis by a single immunocyte, or by a clone, of more than one class of heavy (H) chain is not in contrast with the principle of the uniformity of the antibody-combining site because the constant-region H-chain genes share the same set of V. genes. It is possible, therefore, that once a given V. gene has been chosen by a cell it may become associated with more than one constant-region gene on the same chromosome by successive translocations of the VH gene itself (Gally and Edelman 1970) or by the successive excision of the DNA region between Vs and the different C. genes (Tonegawa et al., this volume). These models predict that different H-chain classes may be synthesized by one clone of B immunocytes, a phenomenon for which there is ample experimental evidence; but if only one chromosome is active, a single cell should not simultaneously synthesize more than one class (except of course for a period corresponding to the persistence and function of previously synthesized mrna). This problem concerning the genetic control of immunoglobulin synthesis was one reason why we were led to consider the evidence for the simultaneous synthesis of more than one H chain by single lymphocytes; another reason was that if two classes of immunoglobulins are regularly synthesized by single lymphocytes, particularly as membrane receptors, this probably implies an important physiological function for the simultaneous presence of both classes of receptors. In fact, there are reports in the literature (Takahashi et al. 1969; Litwin et al. 1973; van Boxel and Buell 1974) of synthesis of more than one H chain by single cells in human lymphoblastoid lines cultured in vitro. These observations, however, might be explained on the basis of the chromosomal abnormalities often found in these cells (see, for instance, Takahashi et al. 1969) and, although interesting in themselves, might not be very meaningful with respect to the biological problems indicated above. One clear case of the synthesis of multiple classes of immunoglobulins by single immunocytes being explained by an abnormal choromosomal set is in lines obtained by the fusion of different plasmacytoma cells between themselves or with normal immunocytes (K6hler and Milstein 1975). Of course in this case, the restrictions pertaining to the synthesis of one L-chain type and one H-chain variable region are also abrogated. With this in mind, we will consider in this report mainly the evidence for multiple H-chain synthesis by single normal lymphocytes. The evidence from lymphoblastoid lines or from lymphomas should always be considered with a view to possible chromosomal abnormalities due to cell fusion or chromosome nondisjunction. It is interesting to point out, however, that the latter case would still be compatible with the synthesis of different H chains with the same variable region (idiotype); the finding of only one L chain would not clearly disprove either cell fusion or chromosome nondisjunction. The most common associations of two different H chains detected in single normal immunocytes are g plus 8 (Rowe et al. 1973; Knapp et al. 1973) and g plus 7 chains (Pernis et al. 1971). In this report, we will consider the evidence in favor of an actual simultaneous synthesis of these chains by the cells. EXPERIMENTAL DETAILS Experiments were carried out in vitro on human peripheral blood lymphocytes and on mouse and rabbit spleen and lymph node lymphocytes. Periph- 175

2 176 PERNIS, FORNI AND LUZZATI eral blood lymphocytes from healthy adults and from cord blood were separated on Ficoll-Urovison mixtures as described previously (Rowe et al. 1973). Mouse and rabbit lymphocyte suspensions were prepared by teasing in cold balanced-salt solution (BSS). Mouse spleen cultures were performed either in Falcon Microtest II plates or in Linbro FB-16-24TC plates, in the conditions described by Kearney and Lawton (1975). A human continuous lymphoblastoid line was kept in culture in RPMI medium supplemented with fetal calf serum. Antisera to human immunoglobulin classes were as described by Rowe et al. (1973). Antisera to rabbit IgM and IgG, as well as anti-immunoglobulin allotype antisera, were the same as used and described previously (Pernis et al. 1970, 1971). Antisera to mouse IgM and IgG subclasses were raised against purified mouse myeloma proteins and absorbed with Sepharose-coupled, purified myeloma immunoglobulins of the other classes as well as with F(ab')2 fragment prepared from normal, pooled mouse IgG by pepsin digestion. The specificities of the antisera were routinely assessed by double immunofluorescent staining of mouse spleen cell smears with the fluorochrome-conjugated antibodies in all possible combinations as well as on plasmacytoma cells of the different classes. The immunoglobulin fractions of the antisera prepared by repeated precipitation with 1.6 M ammonium sulfate and DEAE-cellulose chromatography were conjugated with fluorescein (FITC) or tetramethylrhodamine (TRITC) isothiocyanate (BBL, Baltimore, Md.), basically according to the method of Cebra and Goldstein (1965), purified by Sephadex G-50 gel filtration and DEAE-cellulose chromatography, and used routinely at an IgG concentration of 0.5 mg/ml. Double staining for surface immunoglobulins was performed by incubating the cells with a fluorescein-conjugated antiserum for 20 minutes in the cold and for 10 minutes at 37~ to allow the reacted immunoglobulin to redistribute at one pole of the cell (capping). After washing with cold medium (BSS containing 10% fetal calf serum and 10 mm sodium azide), the cells were incubated for 30 minutes in ice with a rhodamine-conjugated antiserum of different specificity. The cells were then smeared on slides, fixed briefly in ethanol, rehydrated, and mounted in phosphate-buffered glycerol. For detection of intracytoplasmic immunoglobulins, cells were flattened on slides with a Shandon-Elliott cytocentrifuge and fixed in cold ethanol for 10 minutes. After washing in phosphatebuffered saline (PBS), the cells were reacted for 30 minutes in a moist chamber at room temperature with one antiserum, thoroughly washed with PBS, and restained with a second antiserum of different specificity, coupled with a different fluor- chrome, for 30 minutes, rinsed in PBS, and mounted in glycerol. For the simultaneous detection of surface and intracytoplasmic immunoglobulins, the cell suspensions were first stained in the cold and in sodium azide-containing medium with a rhodamineconjugated antiserum, washed, smeared on slides by means of a cytocentrifuge, fixed in ethanol, and processed for intracytoplasmic staining by a fluorescein-labeled antiserum of the same or of a different specificity as indicated above. A Leitz Orthoplan, equipped with an Opak-Fluor vertical illuminator and a HBO-100W/2 mercury vapor lamp (Osram), and a Zeiss Photomikroskop II, also equipped with a vertical illuminator and a HBO200 mercury vapor lamp, have been used for the fluorescence microscopy. Both microscopes were equipped with combinations of filters and dichroic mirrors for the selective visualization of fluorescein or rhodamine. Counts of total cells were made by phase-contrast microscopy. RESULTS AND DISCUSSION IgM Plus IgD Normal B lymphocytes. The majority of normal B lymphocytes in the peripheral blood and in the lymphoid tissues of different species have two classes of membrane immunoglobulins, namely IgM and IgD (Rowe et al. 1973; Knapp et al. 1973; Melcher et al. 1974; Abney and Parkhouse 1974; Pernis et al. 1975). There is evidence that both classes are products of the cell that carries them and that both can reappear on the membrane after selective removal by "capping" with specific antisera or with anti-fab antibodies that react with both classes of immunoglobulins. Both IgM and IgD have also been seen to reappear after removal from the membrane with proteolytic enzymes (Ferrarini et al. 1976b), although in these conditions IgD may fail to reappear if the cells have been treated with an excessive concentration of the enzyme. The simultaneous presence on the membrane of a given lymphocyte of both ~ and 8 chains that have been produced by that cell does not necessarily imply simultaneous synthesis of both chains, nor, in particular, simultaneous synthesis of both their mrna molecules, since a "switch" from the synthesis of one mrna to the synthesis of another, combined with a relatively long life of the mrna molecules, might explain the observation. In fact, both a 8 to ~ and a ~ to 8 switch have been considered (Rowe et al. 1973; Vitetta and Uhr 1975). However, there is now good, albeit indirect, evidence that the B lymphocytes are first generated in the bone marrow or early in ontogeny as cells carrying ~ chains only (Vitetta et al. 1975), and that later in their development they acquire 8 chains as part of a process of antigen-independent matura-

3 ONE LYMPHOCYTE, TWO IMMUNOGLOBULINS 177 tion. On the other hand, at least some of these lymphocytes that have/~ and 8 on their mem.branes never lose the ability to synthesize ~ chains and can mature, with progressive loss of membrane 8 chains, to the stage of an IgM-secreting plasma cell. This process is clearly detectable in the bone marrow of patients with Waldenstr6m macroglobulinemia (Pernis et al. 1974), where practically all B immunocytes are members of one clone that includes lymphocytes with membrane ~ plus 8 chains as well as plasma cells that have both ~t and 8 or only on their membranes and that contain IgM in the cytoplasm, which is the monoclonal protein that is secreted. A percentage of IgM-containing plasma cells in human tonsils also have 8 chains on their membranes, indicating their possible origin from lymphocytes carrying membrane ~ and 8 (Ferrarini et al. 1976a). Therefore, the possible sequence of events in the maturation of at least some of the B-lymphocyte clones is that they begin with the synthesis of ~ chains only, then they activate the synthesis of 8 chains, without inactivating that of chains, and finally mature to IgM-secreting plasma cells that have lost their ability to synthesize 8 chains. This obviously implies that there should be lymphocytes capable of simultaneous synthesis of and 8 mrna molecules, in keeping with the very high proportion of lymphocytes showing both chains. The alternative possibility of a/~ to 8 switch followed by a reversal (8 to #) does not make sense either in molecular or in physiological terms. However, this conclusion is based on the somewhat indirect evidence indicated above for a maturation of ~ plus 8 lymphocytes to IgM-secreting plasma cells; it would be desirable to obtain more direct support for this process, possibly with the use of a cell sorter. Monoclonal lymphomas. Several cases of chronic lymphocytic leukemia have been described in which practically all the lymphocytes in the peripheral blood that had membrane immunoglobulins showed both /~ and 8 chains, in analogy with the above described findings on the majority of normal peripheral blood lymphocytes. Particularly interesting were those cases in which it could be shown that both IgM and IgD were marked by the same idiotype (Fu et al. 1974). It is difficult to explain on the basis of a/~ to 8 or 8 to ~ switch the fact that in these cases practically all the cells carry both chains, since clearly one should find at least a proportion of cells having one chain only. Again the more probable explanation is simultaneous synthesis, but the evidence in this case too is indirect. Study of a lymphoblastoid line. Having in mind the limitations imposed on the interpretation of such studies by the likely chromosomal abnormalities of these cells, as indicated in the introduction, we chose for this study one human lymphoblastoid line established and cloned in vitro (line BL) in which all the cells examined showed a normal chromosomal pattern. In fact, the chromosomal analysis showed that not only was there the normal number of 46 chromosomes, but also that this number was not the result of obvious deletions of chromosomes in some pairs being compensated for by additions in others (pseudodiploid); so, with the limitation that a chromosomal banding analysis was not done, the cells of this line should have carried only two chromosomes with the DNA region for the H-chain genes. When examined in exponential growth phase, all the cells of this line reacted strongly with both the anti-/~ and anti-8 antisera, but not with antisera directed against -/or a chains. They were all positive with an anti-k antiserum but did not react with an anti-k. In this phase, all the cells in the culture were proliferating since, as shown by autoradiography, they all actively incorporated [3H]thymidine. Under these conditions, if one excludes the unlikely possibility of alternating/~ to 8 and 8 to switches, the only reasonable explanation for the double presence of ~ and 8 chains on all the cells is simultaneous synthesis. This conclusion was further supported by experiments in which the cells were treated with an anti-8 antiserum (diluted 1/50 in the culture medium) for 48 hours. This treatment resulted in the rapid disappearance of IgD from the surfaces of the cells, whereas IgM was unaffected, but IgD reappeared on all the cells within 6 hours after the antiserum was removed. The converse happened in cells treated with an anti-/~ antiserum, but, in this case, the disappearance of IgM from the membrane was never complete. IgM Plus IgG Cells containing antibodies reacting with the same antigen but belonging to two different classes (IgM and IgG) were first detected in rats by Nossal et al. (1964) using the microdroplet technique. In these experiments, the proportion of "double" cells was in the range of 10-14% of all cells containing antibodies, the doubles being more frequent at days 3 and 4 of a secondary response. In mice, 589 days after a primary immunization with sheep erythrocytes, Nordin et al. (1970) detected a smaller (2-4%) proportion of cells producing antibodies belonging to the IgM and IgG classes (only the IgG1 subclass was studied). They concluded that the large number of cells producing IgG1 antibodies appearing during that period could not be originated by a "switch" from cells previously secreting IgM. A small proportion of plasma cells containing both IgM and IgG in their cytoplasm can be demonstrated regularly in rabbit lymphoid tissues by

4 178 PERNIS, FORNI AND LUZZATI immunofluorescence (Table 1). When the same cell populations are scored for cells that have IgG in the cytoplasm and IgM on their membranes, the population of "doubles" is much higher, up to 15-16% of all the IgG-containing cells. These cells had matching a-locus (Vj~) and b-locus (K light chain) allotypes on their membranes and in their cytoplasm (Pernis et al. 1971), in accord with the restrictions imposed by an expected identical antibody-combining site of the membrane IgM and the intracytoplasmic IgG. On the basis of this observation, the possibility was considered that IgG-secreting cells might derive from precursor lymphocytes having IgM on their membranes. This possibility was subsequently established by experiments showing the deletion of IgG production as a consequence of treatment with anti-/z antisera both in vitro (Pierce et al. 1972) and in vivo (Kincade et al. 1970; Lawton et al. 1972; Manning and Jutila 1972) and also by work done with the fluorescence-activated cell sorter (Jones et al. 1974). The "switch" within a clone from/z-chain synthesis to 7-chain synthesis is therefore established as the normal process that originates cells secreting IgG, and this easily explains the finding of cells that simultaneously bear IgM and IgG. It is not clear, however, whether this switch may include a period in which a single cell simultaneously synthesizes /Z and 7 chains and, in particular,/z and7 mrnas. We first thought of this latter possibility when we observed that out of 20 human IgG myelomas, there was one (unpubl. obs.) in which practically all the IgG-containing cells in the bone marrow showed membrane IgM; this pattern had not changed in a second bone marrow sample taken two months later. Another human IgG myeloma with this property has been observed by Preud'homme and Seligmann (1974), but since in neither case has a chromosomal study been performed, the conclusions that can be drawn from these two myelomas are limited. Recently, Kearney and Lawton (1975) have observed that it is possible to induce the appearance of a fair number of IgG-containing cells in cultures of mouse lymphocytes stimulated by gram-negative lipopolysaccharide (LPS) at low cell density. Since the appearance of IgG-containing cells was blocked by the addition of an anti-/z antiserum, it appeared that the system was suitable for the study of the IgM to IgG switch in vitro. In fact, these authors have reported (Kearney et al. 1976) that LPS stimulation of newborn mouse lymphocytes results in a progressive increase of cells bearing IgG2 on their membranes, and that for up to 33 hours of culture, all these cells also have membrane IgM, whereas cells with both [gm and IgG in the cytoplasm were quite rare at all times. In experiments with newborn cells, the proportion of cells double on the membrane for IgM and IgG fell rapidly after 33 hours of culture, whereas with cells from adult mice, the "doubles" were more persistent. In fact, it was still possible to suppress substantially the appearance of most cells with intracytoplasmic IgG2 by the addition of anti-/z antibodies as late as day 4 of culture. We have performed similar experiments in which the three main IgG subclasses, namely, IgG1, IgG2a, and IgG2b, were studied separately and have extended these studies to the very early (6 hr culture) changes that take place after addition of LPS. Our results, obtained with spleen lymphocytes from adult mice, differ in some respects from those of Kearney et al. (1976), who did most of their work on cells from newborn animals. We have studied all the possible combinations of IgM and IgG subclasses (a) on the membranes of the cells, (b) on the membrane versus the cytoplasm, and (c) in the cytoplasm at various times after addition of LPS. lgm and IgG on the lymphocyte membrane. We first studied the effect of LPS on membrane immunoglobulins in short-term experiments. Table 2 shows the results of one experiment performed by incubating mouse spleen cells with LPS and staining for membrane immunoglobulins after 6 hours of culture. Table 1. Simultaneous Presence of IgM and IgG in Rabbit Plasma Cells Percent of IgG-containing cells Day after ]gg containing Tissue Immunization stimulation cells with membrane IgM cells containing both IgM and IgG Spleen SRBC primary 7 8 SRBC secondary Lymph node SRBC secondary 5 ferritin secondary

5 ONE LYMPHOCYTE, TWO IMMUNOGLOBULINS 179 Table 2. Early Appearance of Surface IgG on LPS-stimulated Mouse Spleen Lymphocytes Cells with surface IgG (percent of total cells in culture) Culture IgM conditions without IgG IgM IgG1 IgG1 IgM + IgG2a IgG2a IgM + IgG2b IgG2b No LPS time No LPS 6 hr LPS 6 hr Several elements of Table 2 deserve comment: 1. The spleen lymphocytes of our adult CBA/JCr mice have a relatively high number of lymphocytes with membrane immunoglobulins, and, among them about one out of six has IgG of one subclass or another. It is unlikely that these IgG were passively adsorbed or that the relevant conjugates were bound nonspecifically through the Fc receptor since no cells were seen that reacted with more than one of the reagents directed against the different IgG subclasses. Many lymphocytes with membrane IgG also had membrane IgM. 2. Six hours of culture with LPS resulted in a considerable increase in the proportion of lymphocytes with membrane IgG. This increase is practically all due to an increase in the number of cells that carry both IgM and IgG on the membrane, whereas there is no significant increase in the percentage of lymphocytes that carry IgG without IgM. Also, the increase of the membrane IgG population does not generate any cell with membrane IgG of two different subclasses (many hundreds of cells were scored in each double staining condition for different IgG subclasses). No significant changes in the distribution of lymphocytes with different membrane immunoglobulins was observed in cultures maintained for 6 hours without LPS. 3. In the LPS-stimulated cultures, there is some increase in the total percentage of cells with membrane immunoglobulins, but the majority of the increase of cells with membrane IgG (plus IgM) seems to be due to the appearance of IgG molecules on cells that previously showed IgM without IgG, since there is an almost proportional decrease of this latter population. Experiments are now in progress to investigate whether the proportion of cells with membrane IgM without IgG that persists after LPS treatment corresponds to the proportion of cells that carry membrane IgM and IgD. The experiment was repeated three times with comparable results: the membrane IgG population after six hours of culture with LPS was always more than double that of the starting level. In a fourth experiment (Table 3), we studied the effect of the addition of an anti-/~ antiserum on the rapid appearance of membrane IgG induced by LPS. As shown in Table 3, the presence of anti-ft antibodies at sufficient concentration not only blocked the increase in percentage of cells with membrane IgG (plus IgM) induced by LPS, but actually caused the disappearance of membrane IgG from those IgM plus IgG cells that were present in the original sample, while apparently not affecting those cells that had membrane IgG without IgM. As could be predicted, the anti-re antiserum also caused the loss of membrane IgM by all lymphocytes. We are now studying this phenomenon further, including membrane IgD in the pattern, and comparing the effect of anti-immunoglobulin antisera in the presence of LPS with what happens in the absence of the mitogen. It is worth remembering that it has been established (Andersson et al. 1974) that anti-immunoglobulin antibodies block the immunoglobulin Table 3. Effect of Anti-re Antiserum on the Expression of Membrane Immunoglobulins in the Presence of LPS Cells with surface IgG (percent of total cells in culture) Culture IgM without conditions IgG IgM + IgG1 IgG1 IgM + IgG2a IgG2a IgM + IgG2b IgG2b No LPS LPS+ NR IgG a LPS + R anti-re h (250 ~g IgG/ml) LPS + R anti-ft" (50 ftg IgG/ml) a NR IgG = normal rabbit IgG at 250 pg/ml. ~' R anti-/~ = rabbit anti-mouse ~t chain, IgG fraction.

6 180 PERNIS, FORNI AND LUZZATI IgG 1 IgG2a IgG2b ~j101- o.,j *"O'"'"" D r / / 5 5 1/ --/ 3--0 ~ DAYS OF CULTURE i/\'-.. i o Figure 1. Surface immunoglobulins on IgGbearing lymphoid cells in the first 3 days of stimulation with LPS in vitro. (@) Total IgG-bearing cells; (&) cells bearing both IgG and IgM; IO) cells bearing IgG without IgM. synthesis induced by LPS only if the cells are simultaneously exposed to the antiserum and the mitogen. In other experiments, we followed the changes of cells with membrane IgM and IgG throughout a period of three days of culture with LPS. From the results of a typical experiment (Fig. 1), it appears that the cells with membrane IgM plus IgG that quickly increase in percentage after addition of LPS tend to progressively lose membrane IgM and become pure IgG; in fact, at day 3, the proportion of doubles and of cells with IgG only are almost equal, whereas, at the beginning, the doubles are much more frequent. Of course, caution is needed in the interpretation of the variations observed in terms of changes in the expression of membrane immunoglobulins by single cells, since the total cell population is undergoing considerable transformation due to cell death and, after the first 48 hours, also to cell proliferation (see Fig. 2). It is certainly possible that these two processes involve different groups of cells to different extents and therefore contribute to the observed O Q. _z LL 0 o ~' O9 _1 _1 IJJ o _1 +LPS :*,,,~ LPS n J.;s A.~,~ "~ -LPS ~ 2'4 ' HOURS OF CULTLIRE A ~9.J _J W (..) _J s 0 X if) I.-- ff) Figure 2. Changes in cell population during the first 3 days of stimulation of mouse spleen cells with LPS in vitro. _.1 m changes in distribution; however, this can only have a small effect on the changes observed after 6 hours of culture since at this time there is certainly no cell proliferation and very limited cell death. Membrane IgM and intracytoplasmic IgG. Cells containing intracytoplasmic immunoglobulins first appeared in our LPS-stimulated cultures at day 3 and progressively increased in percentage until (at day 5 or 6 in different experiments) about 60% to 70% of all blasts had immunoglobulins in their cytoplasm. IgM-containing cells appeared first and were the predominant population throughout; cells containing IgG of the different subclasses first appeared at day 4 and then increased in frequency up to about 20% of all ceils at day 7 of culture. It was of interest to study how many of these latter cells had membrane IgM, in analogy with the IgG-containing cells with membrane IgM that we detected in rabbit lymphoid tissues (Pernis et al. 1971, and Table 1). Figure 3, summarizing the results of this study, indicates that a high proportion of IgG-containing cells of the different subclasses have membrane IgM at day 5 of culture, and that this proportion declines in the following days. In principle, this re- "l- F- -,.I,,,~. Off) )1oo (.9< Z ~50-_ ZO o IgG1 IgG2a IgG2b rl DAYS OF CULTURE Figure 3. Presence of IgM on the membrane of IgG-containing cells in LPS-stimulated cultures of mouse spleen cells.

7 ONE LYMPHOCYTE, TWO IMMUNOGLOBULINS 181 sult is comparable to that reported by Kearney et al. (1976) for similar experiments on adult mouse lymphocytes, although the decline of the "double" population appears to take place earlier in their cultures. Cytoplasmic IgM and cytoplasmic IgG. Cells that had been cultured for 5 to 7 days in the presence of LPS were flattened on cytocentrifuge slides and, after fixation, subjected to double staining to detect the presence of more than one class of immunoglobulin in the cytoplasm. The results are reported in Tables 4 and 5. It appears that at day 5 there are many cells that have both IgG and IgM in their cytoplasm (as many as those that at the same day have intracytoplasmic IgG and membrane IgM), but that not one single IgG-containing plasma cell (out of many hundreds scored) had in the cytoplasm two different IgG subclasses. At day 7, the total percentage of IgG-containing cells has increased, but the aliquot that has both IgG and IgM in the cytoplasm is drastically reduced (to about 10% of the IgG-containing cells and 2-3% of total cells). These observations were repeated in seven different experiments. The high proportion of "double" cells observed by intracytoplasmic staining at day 5 was surprising in view of the fact that these cells are rare in lymphoid tissues, even after stimulation (see Table 1), and has not been observed by Kearney et al. (1976) in LPS-stimulated cultures of newborn spleen lymphocytes. The possibility that one or more of our antiimmunoglobulin reagents lacked the required class specificity seems unlikely not only in view of the controls performed (see Experimental Details), but also because of the complete absence of doubly stained cells for different IgG subclasses and, above all, considering the results with the cells at day 7. Table 4. IgM- and IgG-containing Cells in LPS-stimulated Mouse Spleen (Day 5) Percent of total plasma cells of a given class IgG class pure + lgm IgG1 + IgG2a + IgG2b IgG IgG2a IgG2b The discrepancy with the observations on lymphoid tissues might be due to a different behavior of cells in LPS-stimulated cultures or might perhaps be explained by the possibility that cells containing both IgM and IgG are also frequent in vivo but are located in some areas of the tissues (e.g., in the germinal centers) from which they are difficult to bring into single cell suspension by the ordinary teasing procedures without causing too much mechanical damage. In fact, most of our doubly stained cells had the appearance of blasts or medium-size lymphocytes, while very few of them looked like mature plasma cells. The discrepancy with the observations of Kearney and Lawton (1976) is more difficult to explain: Perhaps it was due to the fact that we counted all the cells that showed what we considered a significant staining and not only the bright ones, or it might be due to the use of adult mouse spleen cells in which the loss of the synthesis of IgM in cells that have started the synthesis of IgG appears to be more delayed than in cells from newborns. On the whole, all our work on the simultaneous presence of IgM and IgG in single cells from LPSstimulated cultures indicates that these cells appear very early after stimulation and that they persist for 4-5 days before they eventually mature to cells that secrete only one class. We do not know what the ultimate development of the cells derived from the lymphocytes with membrane IgM plus IgG is, nor, in particular, whether they can generate only IgG plasma cells or IgM plasma cells as well. This problem, as well as the problem of those cells that proliferate in the cultures and have membrane immunoglobulins but remain negative for intracytoplasmic staining, remains to be investigated. With regard to the question of simultaneous synthesis of ~ and T chains by single lymphocytes, we think that the existence of numerous cells with beth IgM and IgG on their membranes already after 6 hours of stimulation with LPS, together with the persistence of these cells for some days and the appearance of cells, presumably derived from them, that have cytoplasmic IgG and membrane IgM or beth IgM and IgG in the cytoplasm (all these conditions being equivalent for considering the simultaneous expression of ~ and T), indicates the possibility of simultaneous synthesis of/~ and T chains by a single cell. It is therefore conceivable that the Table 5. IgM- and IgG-containing Cells in LPS-stimulated Mouse Spleen Cultures Percent of total cells in the culture Total positive IgM without Day of culture plasma cells IgG ]gm + IgG1 IgG1 IgM ~- IgG2a IgG2a IgM + IgG2b IgG2b Values given are average of two experiments.

8 182 PERNIS, FORNI AND LUZZATI IgM plus IgG condition is not basically different from the IgM plus IgD condition: that is, in both cases there is simultaneous synthesis of two different heavy chains by the same cell, although the period of simultaneous synthesis appears to be longer for IgM plus IgD than for IgM plus IgG. It is possible that in both cases the simultaneous synthesis of two different chains reflects the simultaneous synthesis of the two corresponding mrna molecules, although this has not been proved. The next question is whether there are cells that synthesize/~ plus ~ plus T or 8 plus T. We are presently investigating this problem in the mouse. In man, the evidence so far available for normal immunocytes indicates complete reciprocal exclusion of 8 and T (Ferrarini et al. 1976a), and therefore no evidence has been found for the ability of a single, normal immunocyte to synthesize simultaneously more than two heavy chains. It appears that the/x plus 8 and the/~ plus T populations are separate, although a passage of a given cell from the former to the latter group is possible through a sharp 8 to T switch (for which there is indirect evidence, Pernis 1975) with a rapid loss of 8. It is also quite clear from our results that there is complete exclusion among the different IgG subclasses, and thus, for the moment, one may put forward the empirical rule that normal B immunocytes can simultaneously synthesize two, but no more than two, different immunoglobulin heavy chains, one of which must be/x. This rule is provisional and does not take into account the a chains, whose cellular expression has not been sufficiently studied. What meaning the simultaneous synthesis of two different heavy chains by single immunocytes has for the genetic control of immunoglobulin synthesis and for the physiology of B immunocytes remains to be studied. Acknowledgment We are grateful to Dr. V. Miggiano for the chromosomal analysis and autoradiographic studies of cells of the BL lymphoblastoid line. REFERENCES ABNEY, E. R. and R. M. E. PARKHOUSE Candidate for immunoglobulin D present on murine B lymphocytes. Nature 252: 600. ANDERSSON, J., W. W. BULLOCK and F. MELCHERS Inhibition of mitogenic stimulation of mouse lymphocytes by anti-monse immunoglobulin antibodies. I. Mode of action. Eur. J. lmmunol. 4: 715. CEBRA, J. J. and G. GOLDS'rEIN Chromatographic purification oftetramethylrhodamine-immune globulin conjugates and their use in the cellular localization of rabbit globulin polypeptide chains. J. Immunol. 95: 230. FERRAmNI, M., G. VIALE, A. RISSO and B. PERNm. 1976a. A study of the immunoglobulin classes present on the membrane and in the cytoplasm of human tonsil plasma cells. Eur. J. Immunol. 6: 562. FERRARINI, M., G. CORTE, G. VIALE, M. I,. DURANTE and T. BARGELLESI. 1976b. Membrane Ig on human lymphocytes: Rate of turnover of IgD and IgM on the surface of human tonsil cells. Eur. J. Immunol. 6: 372. Fu, S. M., R. J. WINCHESTER, T. FELZI, P. D. WALZER and H. G. KUNKEL [diotype specificity of surface immunogtobuiin and the maturation of leukemic bone marrow-derived lymphocytes. Proc. Nat. Acad. Sci. 71: GALLY, J. A. and G. M. EDELMAN Somatic trans]ocation of antibody genes. Nature 227: 341. JONES, P. P., H. TACIER-EUGSTER and L. A. HERZENBERG Lymphocyte commitment to Ig allotype and class. Ann. Immunol. (Inst. Pasteur) 125C: 271. KEARNEY, J. F. and A. R. LAWTON B lymphocyte differentiation induced by lipopolysaccharide. I. Generation of cells synthesizing four major immunoglobulin classes. J. Immunol. 115: 671. KEARNEY, J. F., M. COOPER and A. R. LAWTON B lymphocyte differentiation induced by lipopolysaccharide. IV. Development of immunoglobulin class restriction in precursors of IgG-synthesizing cells. J. Immunol. (in press). KINCADE, P. W., A. R. LAWTON, D. E. BOCKMAN and M. D. COOPER Suppression of immunoglobulin G synthesis as a result of antibody mediated suppression of immunoglobulin M synthesis in chickens. Proc. Nat. Acad. Sci. 67: KNAPP, W., R. L. H. BOLHUIS, J. RADL and W. HIJMANS Independent movement oflgd and IgM molecules on the surface of individual lymphocytes. J. ]mmunol. 111: K~HLER, G. and C. MILSTEIN Continuous cultures of fused cells secreting antibodies of predefined specificity. Nature 256: 495. LAWTON, A. R., R. ASOFSKY, M. B. HYLTON and M. D. COOPER Suppression of immunoglobulin class synthesis in mice. I. Effect of treatment with antibody to/~-chain. J. Exp. Med. 135: 277. LITWIN, S. D., P. K. LIN, T. H. HUETrEROTH and H. CLEVE Multiple heavy chain classes and light chain types on surface of cultured human lymphoid cells. Nature New Biol. 246: 179. MXKEL:4, The specificity of antibodies produced by single cells. Cold Spring Harbor Symp. Quant. Biol. 32: 423. M.~KEIL4, O. and A. M. CROSS The diversity and specialization of immunocytes. Prog. Allergy 14: 145. MANNING, D. D. and J. W. JUTILA Immunosuppression of mice injected with heterologous anti-immunoglobulin heavy chain antisera. J. Exp. Med. 135: MELCHER, U., E. S. VITETTA, M. MCWILLIAMS, M. E. LAMM, J. M. PHILLIPS-QUAGLIATA and J. W. UHR Cell surface imrnunoglobulin. X. Identification of an IgD-like molecule on the surface of murine splenocytes. J. Immunol. 113: NORDIN, A. A., H. COSENZA and S. SELL Immunoglobulin classes of antibody forming cells in mice. II. Class restriction of plaque-forming cells demonstrated by replica plating. J. Immunol. 104: 495. NOSSAL, G.J.V., A. SZENEERG, G. L. ADAand C. M. AUSTIN Single cell studies on 19S antibody production. J. Exp. Med. 119: 485. PERNIS, B Relationships between the heterogeneity of immunoglobulins and the differentiation of plasma cells. Cold Spring Harbor Symp. Quant. Biol. 32: The effect of anti-igd antiserum on antibody production of rhesus monkeys. In Membrane receptors of lymphocytes (ed. M. Seligmann et al.), p. 25. North- Holland, Amsterdam. PERNIS, B., J. C. BROUET and M. SELIGMANN IgD and IgM on the membrane of lymphoid cells in macro-

9 ONE LYMPHOCYTE, TWO IMMUNOGLOBULINS 183 globulinemia. Evidence for identity of membrane IgD and IgM antibody activity in a case with anti- IgG receptors. Eur. J. Immunol. 4: 776. PERNIS, B., L. FORNI and L. AMANTE Immunoglobulin spots on the surface of rabbit lymphecytes. J. Exp. Med. 132: Immunoglobulins as cell receptors. Ann. N.Y. Acad. Sci. 190: 420. P~.RNXS, B., L. FORNX and K. L. KNmHT The problem of an IgD equivalent in non-primates. In Membrane receptors of lymphocytes (ed. M. Seligmann et al.), p. 57. North-Holland, Amsterdam. PIERCE, C. W., S. M. SOLLmAY and R. ASOFSKY Immune response in vitro. IV. Suppression of primary TM, TG and TA plaque-forming cell responses in mouse spleen cell cultures by class specific antibody to mouse immunoglohulins. J. Exp. Med. 135: 675. PR~.UD'HOMME, J. L. and M. SELIGMANN Surface immunoglobulins on human lymphoid cells. In Progress in clinical immunology (ed. R. S. Schwartz), vol. 2, p Grune and Stratton, New York. RAFF, M. C., M. FELDMANN and S. DE PETRIS Monospecificity of bone marrow-derived lymphocytes. J. Exp. Med. 137: ROWE, D. S., K. HUG, L. FORNI and B. PERNIS Immunoglobulin D as a lymphocyte receptor. J. Exp. Med. 138: 965. TAKAHASHI, M., N. TAKAGI, Y. YAGI, G. E. MooRE and D. PRESSMAN Immunoglobulin production in cloned sublines of a human lymphocytoid cell line. J. Immunol. 102: VAN BOXEL, J. A. and D. N. BUELL IgD on cell membranes of human lymphoid cells with multiple immunoglobulin classes. Nature 251: 443. VITETTA, E. S. and J. W. UHR Immunoglobulin receptors revisited. Science 189: 964. VITETTA, E. S., U. MELCHER, M. McWILLIAMS, M. E. LAMM, J. S. PHILLIPS-QUAGLIATA and J. W. UHR Cell surface immunoglohulin. XI. The appearance of an IgD-like molecule during ontogeny. J. Exp. Med. 141: 206.

10 Synthesis of Multiple Immunoglobulin Classes by Single Lymphocytes B. Pernis, L. Forni and A. L. Luzzati Cold Spring Harb Symp Quant Biol : Access the most recent version at doi: /sqb References alerting service This article cites 32 articles, 17 of which can be accessed free at: Article cited in: Receive free alerts when new articles cite this article - sign up in the box at the top right corner of the article or click here To subscribe to Cold Spring Harbor Symposia on Quantitative Biology go to: Copyright 1977 Cold Spring Harbor Laboratory Press