Eur. J. Immunol. 1986.16: 1297-1301 Lifespan of Ig-producing cells 1297

Faith Ho’, Jennifer E. Lortan,

Mahmood Khan cell populations* Distinct short-lived and long-lived antibody-producing

Ian C. M. MacLennan and

Department of Immunology, University of Birmingham, Birmingham This report analyzes the life span of Ig-containing cells (IgCC) in different sites of

antibody production. The experimental approach was based upon the observations that most’IgCC are derived from proliferating precursors while-IgCC themselves are mainly nondividing end cells. Rats were given a continuous infusion of [3H] thymidine via an osmotic pump inserted in the peritoneal cavity. At intervals of 1 , 3 , 5 or 10 days after starting infusions, tissues were taken and analyzed by a combination of immunohistology and autoradiography to identify the proportions of IgCC which had gone through S phase of the cell cycle during the period of infusion. After 3 days infusion the median and (range) percent-labeled IgCC in the medullary cords of mesenteric and cervical lymph nodes and the red pulp of the spleen were, respcc- tively, 88 (81-90), 75 (66-77) and 88 (82-93). Conversely that for IgCC in bone marrow was only 13 (11-17) and that in the lamina propria of the jejunum 47 (33-68). The rate of increase in labeling of bone marrow IgCC with length of infusion was approximately linear. Extrapolation of this slope suggests that bone marrow IgCC have a life span in excess of 3 weeks. The slopes of increase in IgCC labeled with time for lymph nodes and spleen were clearly biphasic suggesting that while most IgCC in these tissues have a life span of less than 3 days, there is also a minor population of long-lived IgCC. The lamina propria appears to have approximately equal propor- tions of long and short-lived IgCC. The life span of IgCC, with the exception of IgMCC, appears to be a feature of the site of antibody production rather than the Ig class produced. Almost all IgM-containing cells were found to be short lived.

1 Introduction

Antibody-secreting cells (ASC) are found in a variety of sites in the body. Their location relates to three interdependent factors: (a) the site of antigen-driven activation of B cells; (b) the nature of the antigen, and (c) the stage of the immune response [ 1-41. For example, primary T cell-dependent B cell activation in the spleen and peripheral lymph nodes (LN) usu- ally gives rise, respectively, to ASC in the red pulp of the spleen and the medulla of the same LN [3]. By contrast, in established secondary responses, B cells activated in the spleen and peripheral LN gives rise to ASC in bone marrow (BM) [3]. Antibody production in the lamina propria of the gut mainly follows B cell activation in mesenteric LN [5-71.

The kinetics of antibody responses depends upon a number of factors including the number of B cells activated by antigen, the extent of proliferation of B blasts and the rate of B blast differentiation to ASC. The active life span of ASC must also be of importance. Studies of ASC life span have been reported for ASC in popliteal LN [S] and the lamina propria of the gut [9]. These studies suggest that the active secretory life span of

[I 55691 ~~~~~~~~~

* This work was supported by a grant from the Medical Research Council, U.K. Dr Ho, a visiting research fellow from the Department of Pathol- ogy, University of Hong Kong, was supported by a Royal Society fellowship.

Correspondence: Ian C. M. MacLennan, Department of Immunology, University of Birmingham Medical School, Vincent Drive, Birming- ham B15 2TJ, GB

Abbreviations: IgCC: Immunoglobulin-containing cell ASC: Anti- body-secreting cell LN: Lymph node BM: Bone marrow r3H]dTbd: Tritiated thymidine

most ASC is less than 4 days although a minority of ASC particularly in the lamina propria of the gut may survive for up to 6 weeks.

The object of the present study has been to determine the life span of ASC in the main sites of antibody production and to see if the isotype of antibody produced is related to life span. The experimental approach is based on the observations that most ASC are derived from proliferating precursors but that ASC in general are nondividing end cells [8, 101. Taking advantage of these properties of the majority of ASC it is possible to study the life span of these cells. We have done this by assessing the proportion of Ig-containing cells (IgCC) labeled in rats after varying lengths of [3H]thymidine (PHIdThd) infusion. The results of this study indicate that there are distinct short- and long-lived ASC populations which appear to be generated by B cell activation in different condi- tions.

2 Materials and methods

2.1 Rats

(DA x LouC) F, hybrid rats aged between 9 and 13 weeks were used in the experiments described. They were bred and maintained in the Department of Immunology, University of Birmingham, under standard laboratory conditions. Experi- mental groups were age, sex and weight matched.

2.2 Antibodies

Rabbit anti-rat Ig; pig anti-rabbit Ig and soluble immune com- plex of horse radish peroxidase with rabbit anti-horse radish peroxidase were purchased from DAKO, High Wycombe, GB. Rabbit antibodies to rat p, a , yl, y2a, Y2b and ~2~ heavy

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1986 0014-2980/86/1010-1297$02.50/0

1298 F. Ho, J. E. Lortan, I. C. M. MacLennan and M. Khan Eur. J . Immunol. 1986.16: 1297-1301

chains, generously provided by Dr. H. Bazin, were prepared and their specificity was tested as described by Bazin et al. [111.

2.3 I n vivo [3H]dThd labeling

[3H]dThd, spec. act. 5000 Ci/mmol = 185 TBqlmmol, (code TRA-61 Amersham, GB), was administered as a continuous infusion at 0.5 pCi/g body weightlday via an osmotic pump. The pumps (Alzet osmotic pumps Model 2ML1, Scientific Marketing Associated, London, GB) were loaded according to the manufacturers instructions and inserted into the left iliac fossa of the peritoneal cavity under ether anesthesia. Groups of rats were killed at intervals of 1, 3, 5 and 10 days after starting infusions.

2.4 Histological processing of tissues and autoradiography

Tissues were fixed in formol sublimate for 18 to 24 h followed by dehydration and paraffin wax embedding [12]. Sections of 5p were subjected to immunohistological staining to reveal cytoplasmic Ig heavy chains of the isotypes indicated above using the peroxidase anti-peroxidase technique [13]. Immunoperoxidase-stained tissue sections were dipped in Ilford K-5 emulsion (Ilford Mobberley, GB) and exposed for 5 weeks at - 20°C. The autoradiographs were developed for 3 min in Kodak D19 (Eastman Kodak, Rochester, N Y) developer at 20 "C, fixed and counterstained with Mayer's hemalum.

2.5 Quantitative histology

IgCC were enumerated in the splenic red pulp, the medullary cords of cervical and mesenteric LN, jejunal lamina propria and femoral BM. Sections were examined using a Leitz micro- scope (Wetzlar, FRG) with 10 x eye pieces and a 40 x objec- tive. The number of cells containing antibody of a particular isotype in unit volumes of the tissues specified was determined using the point counting method of Weibel [14]. An 1 cm square grid graticule with 1 mm subsquares was placed in one eyepiece. Sections were examined by serial left to right sweeps from top to bottom of the section until 1000 line intercepts had fallen on the specified tissue area or the entire section had been counted. Results are expressed as IgCC per 1000 line intercepts. The proportion of [3H]dThd-labeled IgCC was assessed by counting 200 anti-Ig-stained cells in a specified area or all the stained cells in those areas of the section if less than 200. Again serial sweeps were made when examining each section. Slides were coded before quantitation so that the person carrying out the counts was unaware of the Ig isotype revealed in the section.

3 Results

3.1 Estimation of the renewal rates of IgCC in different tissues

The proportions of IgCC labeled after different periods of continuous [3H]dThd infusion in vivo was estimated by deter- mining the proportion of IgCC of each Ig heavy chain isotype

Table 1. Percent Ig-containing cells labeled after different lengths of [3H]dThd infusion")

Ig isotype: Total Ig P a Y l Y-7 Y?h Y2c

Days ["HJdThd: 1 3 5 10 3 10 3 10 3 10 3 10 3 10 3 10 Tissue

Spleen red pulp 32 90 89 93 79 92 60 90 32 93 32 86 96 84 91 94 62 73 31* 65 42 81 93 91 54 25 50 90 82 92 91 94 78 77 36 91

Median 37 88 91 92 91 94 57 77 32 91

Cervical LN medulla 66 79 96 94 95 69 97 23 75 21 77 77 96 92 88 65 91 65 81 19 75 74 92 62 27 29 67 89 70 77 50 92 36 83

Median 21 75 76 96 93 88 64 92 32 81

Mesenteric LN medulla 93 77 97 87 95 80 97 65 75 36 93 83 98 88 88 68 99 85 81 26 82 85 98 77 88 39 83 80 YO 77 77 88 98 87 83

22 21 20 46

22

34 31 22 67

38

86 79 95 94

93 < < 62 92 88 < 94 70 94

< 68 93 89 < 89 93

93 < < 69 93

92 46 < 45 77 82 62 88 47 88

48 44 85 67 91 45 79

85 55 W 35 79

92 78 81 66 86 82 89 82 84 77

87 71 a) Results are expressed as percent of 85 92 88 48 62 immunostained cells tritium

labeled. 36 88 82 97 88 88 79 98 86 81 90 86 88 82 69 77 < = fewer than immunostained

8 47 62 84 < < 47 87 < 100* < 88* < < < < * = fewer than 2 0 2 5 immuno-

6 68 61 83 < 100 64 84 55 75 < 71' < < < < No entry = technical failure.

Lamina propria jejunum 8 33 53 88 < < 66 88 SO' < < 67* < < < < cells/slide.

13 46 56 < 42 < < < < stained cells/slide.

Each result shown represents the 8 47 59 84 < < 56 87 75 < 71 < < < < value from a single rat. Animal 3

Median

BM femoral 0 12 34 40 60* 20' 12 38 0* 30 0' 20 < < < < killed at 10 days had a nonfunc- 1 13 37 34 18" 45 8 35 11" 27 0 35 < < < < tioning osmotic pump as assessed 1 17 13 60' IS 4 3 < < by absence of intestinal epithelial 0 11 19 50 20* 42 11 47 13' 33 0' 47 < < < < cell labeling other missing values

Median 1 13 27 40 40* 42 12 38 8* 30 0 35 < < < < represent technical failurebf auto- radiographs.

Lifespan of Ig-producing cells 1299 Eur. J. Immunol. 1986.16: 1297-1301

w 5 m

m : !? 100 a _ > - + u O Y m a - w 80 clr Y d

Y U

u v Y - a " 7 2

z o - t z I L

- - 60

0 2 40

2 0 20 z r o z " W

u L J " a o w - a S C M L B S C M L E S C M L B S C M L E S C U L B S C M L E SCMLB

TOTAL IG IGM IGA I G G ~ IGGPA IGG.?B IGGZC 115569.21 1 3 5 10

D A Y S O F l3H1 dThd INFUSION Figure 2. The percent of IgCC containing different Ig classes and sub- classes in different tissues. The proportion of these cells labeled after 3 days continuous [3H]dThd infusion in vivo is shown by the hatched area in each column. S = spleen red pulp; C = cervical LN medullary cords; M = mesenteric LN medullary cords; L = lamina propria jejunum; B = BM femur.

- Figure 1 . The median percentages of IgCC in different sites labeled with [3H]dThd after 1, 3, 5 or 10-day continuous [3H]dThd infusion in vivo. SPRP = Splenic red pulp; MLN = medullary cords of mesenteric LN; CLN = medullary cords of cercival LN; LP = lamina propria of jejunum: The results shown are for total IgCC at each site.

which were labeled after 1, 3, 5 or 10 days [3H]dThd infusion. The results are shown in Table 1 and the median percent of labeled IgCC at each site is plotted in Fig. 1. The rate of labeling of IgCC in different sites varies markedly; there are three main patterns: (a) in the red pulp of the spleen, and the medullary cords of both cervical and mesenteric LN, more than three quarters of all IgCC were labeled by 3 days and over 90% at 10 days. The rate of increase in the proportion of

labeled IgCC is approximately linear in these sites to 3 days. After this the rate slows markedly. This pattern of labeling is broadly similar for IgCC of each Ig isotype. However, the proportions of labeled IgGICC and IgGzaCC are lower at 3 days in cervical LN and spleen than those of IgACC and IgMCC. (b) In the BM almost no IgCC were labeled at the end of 24 h infusion. However, many other types of myeloid cell were

Table 2. Relative numbers of Ig-containing cells per unit volume of different tissuesa)

Jg isotype

Tissue

Splenic red pulp 395 320 20 25 2 33 80 473 217 59 127 1 26 213 357 135 25 56 5 10 96 838 179 IS S2 4 15 86

Median 434 198 23 54 3 21 91

Medullary cords cervical LN 154 101 72 217 6 20 315 514 54 7 48 5 32 92 247 25 313 131 2 28 114 359 123 813 173 27 215 1228

Median 303 76 193 152 h 30 365

Medullary cords Mcsenteric LN 885 893 57 37 20 113 227 573 945 60 12 15 67 154 347 1043 62 16 72 315 465

Median 573 945 60 16 20 113 227

Jejunal lamina propria

Median Femoral BM

Median

3 60.1 13 3 0 0 16 9 741 10 6 0 3 19

25 17

13

8 8 S 8 8

878 868

805 293 155 98 I41 148

174 7 0 0 181 44 3 9 0 56

2 9 5 0 0 38

58 92 7 4 161 50 12 0 0 62 2 6 1 0 9 4 20 0 ?- 26

27 56 1 1 4-2

Sum of all

isotypes

795 903 588

1103

849

570 660 686

1710

673

2005 1672 1855

1855

623 931

1084 941

936

462 225 112 I75

200

a) The values given are the number of IgCC with each Ig isotype in the volume of tissue covered by 1000 grid intercepts. For details see Sect. 2.5. Note that no ac- count is made of the total volume of tissue in each of the compart- ments studied. Consequently, the values given reflect the con- centration of IgCC in a particular tissue and not the total number of IgCC in that tissue. Absent values represent technical failure in histological processing of one mesenteric LN.

1300 F. Ho, J. E. Lortan, I. C. M. MacLennan and M. Khan Eur. J. Immunol. 1986.16: 1297-1301

labeled at that stage. After this there was an approximately linear rate of increase in the proportion of labeled IgCC with time, so that on average 40% of IgCC are labeled during 10 days infusion. If this slope is extrapolated, it reaches 100% at approximately 24 days suggesting that most BM IgCC have a life span in excess of 3 weeks. The exception to this is in the small population of IgMCC found in the BM. Forty percent of these cells were labeled during a 3-day infusion. (c) In the lamina propria of the jejunum about half the IgCC were labeled during 3 days [3H]dThd infusion but thereafter the rate of increase in the proportion of labeled IgCC with length of infusion slowed markedly.

3.2 Relation of the renewal rates of IgCC to the relative numbers of IgCC in different tissues

The data given above indicate that there are rapidly labeled and slowly labeled populations of IgCC. The proportions of these populations differ in different tissues. In addition, the relative numbers of cells containing different Ig heavy chain isotypes in each of the sites of antibody production were deter- mined in four normal rats. These results are set out in Table 2. Fig. 2 plots the median number of IgCC for each Ig class and subclass in each tissue with the median number of cells labeled at 3 days in each category. With the exception of IgMCC the proportion of rapidly labeled IgCC appears to be related more to the site of antibody production than the class of antibody produced. Most IgM-containing cells label rapidly irrespective of location. For IgA, IgGl and IgG2aCC both rapidly labeling and slowly labeling IgCC were identified. There were too few IgGzbCC and IgGzcCC detected in the BM and lamina propria to assess whether these were slowly labeling.

4 Discussion

This study suggests that there are distinct short-lived and long- lived populations of IgCC and that the proportions of these in different sites varies markedly. However, a number of factors must be considered which give uncertainty to the method used for determining the life span of IgCC. First the labeling rates of IgCC will not measure life span accurately if IgCC them- selves are dividing. The maximum proportion of IgCC labeled in BM after 24 h infusion was 1%. This implies that the pre- cursors of BM IgCC exit from cell cycle at least 1 day before becoming BM IgCC confirming the results of others that BM IgCC are not in cell cycle [lo, 15, 161. It is possible that the rate of BM IgCC replacement was slowed in the rats studied due to [3H]dThd-induced damage to precursor cells. However, differences between labeling rates of BM compared with spleen and LN IgCC cannot be explained on this basis. Such a toxic effect would be expected to be cumulative and result in a decrease in the rate at which BM IgCC become labeled with time. However, the rate of increase in labeled BM IgCC after the first day was found to be broadly linear. The IgCC life span in BM would also be overestimated if many of these cells were derived from B cell differentiation without proliferation. Rest- ing B cells have been shown to be responsive to B cell matura- tion factors in the absence of B cell proliferation factors [17]. However, maturation without cell division cannot be a major continuous physiological process as the body would rapidly run out of resting B cells if these were not replaced by memory B cell formation or primary B lymphopoiesis.

These considerations seem to make it likely that most BM IgCC have a life span in excess of 3 weeks. This coincides with a lifespan of over two weeks observed for nondividing IgG and IgA but not IgM-secreting cells isolated from human BM [15]. It is less easy to conclude that LN and splenic IgCC have the short life span implied by the rapid rate at which these cells become labeled in vivo. Two factors may result in their life span being underestimated in this study; (a) some of the IgCC may be in cell cycle, and (b) IgCC may leave the LN medulla and splenic red pulp for other sites of antibody secretion. Studies of [3H]dThd uptake by IgCC and ASC in LN suggest that a minority of these cells may be in cell cycle in the first few days following secondary antigenic challenge [8, 101. How- ever, it is possible that these cells are about to exit in efferent lymph without entering the medullary cords. This concept is supported by comparison of the electron microscopic appear- ances of plaque-forming cells in efferent lymph draining from popliteal LN with that of plaque-forming cells within the node. On the fourth day after footpad immunization of rabbits with sheep red cells 93% of plaque-forming cells in efferent lymph did not have well-developed endoplasmic reticulum [ 181, while 95% of plaque-forming cells from the LN themselves had extensive endoplasmic reticulum [19]. These data make it likely that the majority of IgCC in the medullary cords of LN are end cells. If this is the case, the present study appears to show that a high proportion of IgCC in the medullary cords of LN have a life span of less than 3 days. The assessment of IgCC in this study is likely to have identified immediate pre- cursors of ASC in addition to ASC. Results of comparisons of the relationship of IgCC with ASC will depend upon the rela- tive sensitivity of the two assays. Nevertheless, comparisons of IgCC with ASC numbers have generally shown close concord- ance of results [lo, 201.

Benner and his group have extensively studied the sites of antibody production at different stages of antibody responses to different antigens [2,3]. They conclude that a high propor- tion of ASC in BM are derived from T cell-dependent B cell activation in peripheral lymphoid tissue, particularly the spleen, during established immune responses. T cell-depen- dent B cell activation in primary responses or early in second- ary responses, on the other hand, was found to result in anti- body production in the spleen and LN. Combining these con- clusions with the results of the present study, it seems likely that most ASC formed in the period immediately following challenge with T-dependent antigens live less than 3 days while in established antibody responses ASC have a life span in excess of 3 weeks.

In a recent study of virgin B cell recruitment during responses to dinitrophenylated spider crab hemocyanin, it was observed that substantial numbers of virgin B cells recently produced in the BM were activated in the period immediately following antigenic challenge. On the other hand, sustained antibody production was achieved exclusively by activation of memory B cells [20]. Associated studies of the migration patterns of virgin B cells, recently formed in the BM, showed that virgin B cells migrated to extrafollicular sites rich in interdigitating cells but did not reach follicles [22]. It was confirmed in these studies that thoracic duct lymphocytes also passed through the interdigitating cell-rich areas but then progressed to follicles [23]. These data have been used to propose that antigen pre- sented on interdigitating cells is likely to result in T cell-depen- dent activation of both memory and newly formed virgin cells. On the other hand, antigen on follicular dendritic cells seems

Eur. J. Immunol. 1986.16: 1297-1301 Lifespan of Ig-producing cells 1301

likely to be accessible to mature recirculating B cells alone. It seems possible that B cell activation at these two sites may give rise, respectively, to the populations of short and long-lived IgCC described in this study.

The localization of antibody production in responses to thy- mus-independent antigens is heterogenous [4, 241. It is of interest that lipopolysaccharide appears to induce antibody production in BM by direct activation of B cells in that site [24]. It may be that the small number of rapidly labeled IgMCC observed in the BM in the present study were gener- ated in this way.

Received May 8, 1986.

5 References

1 Thorbecke, G. J. and Keuning, F. J., J . Immunol. 1953. 70: 129. 2 Askonas, B. A. and Humphrey, J. H., Biochem. J . 1958. 68: 252. 3 Benner, R., Hijmans, W., Haaijman, J. J., Clin. Exp. Immunol.

4 Koch, G. , Lok, B. D. and Benner, R., Immunobiology 1982.163:

5 Gowans, J . L. and Knight, E. J. , Proc. R. SOC. London Ser. B.

1981. 46: 1.

484.

1964. 159: 257.

6 Hall, J. G. , Parry, D. M. and Smith, M. E. , Cell Tissue Kinet.

7 Parrott, D. M. V. and Furguson, A , , Immunology 1974. 26: 571. 8 Makela, 0. and Nossal, G. J. V., J . Exp. Med. 1962. 115: 231. 9 Mattioli, C. A. and Tomasi, T. B., J . Exp. Med. 1973. 138: 452.

10 Geldof, A. A., Rijnhart, P., van de End, M., Kovs, N. and

11 Bazin, H., Platteau, B., Beckers, A. and Pauwels, R., J . Immunol.

12 Drury, R. A. B. and Wallington, E. A., Carletons Hi~stological

13 Burns, J., Histochemistry 1975. 43: 291. 14 Weibel, E. R. Lab. Invest. 1963. 12: 131. 15 Hibi, T. and Dosch, H.-M., Eur. J. Immunol. 1986. 16: 139. 16 Hibi, T., Chan, M. A, , Petsche, D. and Dosch, H.-M., J .

17 Sidman, C. L. and Marshal, J. D., J . Immunol. 1984. 132: 845. 18 Hummeler, K., Harris, T. N., Harris, S . and Farber, M., J . Exp,

19 Gudat, F. G., Harris, T. N., Harris, S. and Hummeler, K., J . Exp.

20 Walker, L. A. and Johnson, G. D., J . Immunol. Methods 1982.62:

21 Gray, D., MacLennan, I. C. M. and Lane, P. J . L., Eur. J .

22 MacLennan, I. C. M. and Gray, D., Immunol. Rev. 1986. vol. 91,

23 Ford, W. L., Prog. Allergy 1975. 19: 1. 24 Koch, G., Lok, B. D., Van Oudenaren, A. and Benner, R., J .

1971. 5: 269.

Langevoort, H. L., Immunobiology 1984. 166: 296.

1978. 121: 2083.

Technique, Oxford University Press, Oxford 1967.

Immunol. 1986. 136: 3211.

Med. 1972. 135: 491.

Med. 1970. 132: 448.

134.

Immunol. 1986. 16: 641.

in press.

Immunol. 1982. 128: 1497.