Eur. J . Immunol. 1988.18: 1561-1565

John Gordon’, Jennifer A. Cairns’, Michelle J. Millsum’, Steven Gillis’ and Graeme R. GuyA

Department of Immunologyo and BiochemistryA, University of Birmingham, Birmingham and Immunex Corporation+, Seattle

Interleukin 4 and soluble CD23 as progression factors for human B lymphocytes 1561

Interleukin 4 and soluble CD23 as progression factors for human B lymphocytes: analysis of their interactions with agonists of the phosphoinositide “dual pathway” of signalling*

Human B lymphocytes pre-activated for 24 h with a combination of phorbol dibuty- rate [P(BU)~] and ionomycin were found to provide excellent targets for assessing the detailed action of B cell progression factors. Both recombinant interleukin 4 (IL4) and affinity-purified 25-kDa fragment of the CD23 molecule (sol-CD23) were shown to be active in this assay. While the progression activity of IL4 was enhanced by continued co-culture with P(Bu)*, that of sol-CD23 was found to be more strictly dependent upon such a joint application with the phorbol ester. Similar requirements were observed for triggering cell-cycle progression in the pre-activated B cells when using a stimulating CD23 antibody. Ionomycin, in contrast to P(BU)~, did not aug- ment either IL 4 or sol-CD23 in these assays but did enhance significantly the pro- gression activity of an anti-CDw40 antibody. When added to B cells concomitantly with, or prior to, a high dose of phorbol ester, IL 4 unexpectedly down-regulated the subsequent mitogenic response to this agent whereas, when added 24 h later, IL4 up- regulated such stimulations. The latter sequence of additions resulted in a particularly dramatic induction of CD23 at the B cell surface, much more so than seen when B cells were incubated with either IL4 alone or with IL4 and P(Bu)* together. This up- regulation of surface CD23 was, in turn, mirrored by the appearance of large amounts of the soluble form of the molecule in such cultures. The findings are discussed with reference to possible mechanisms through which IL4 and CD23 interact to exert their multiple actions on B cell regulatory pathways.

1 Introduction

B lymphocytes are influenced by multiple regulatory cytokines each seemingly capable of eliciting numerous biological responses [ 11. Conventionally considered to derive principally from T lymphocytes and monocytes, it has become clear that B cells are also capable of producing some of the mediators to which they, themselves, can appropriately respond [2, 31. Fine-tuning within such regulatory networks might depend, at least in part, upon the precise temporal sequence of paracrine and autocrine effects and their interactions with one another. Within this context, a close relationship appears to exist between T cell-derived interleukin 4 (IL4) and the B cell- associated glycoprotein defined by the anti-CD23 cluster of antibodies.

For human B lymphocytes, in both their resting and cycling states, IL4 has been found to be a major inducer of the 45- kDa form of the CD23 molecule at the B cell surface [4, 51. This molecule is a low-affinity receptor for IgE and when cleaved by proteolysis, first to a 35-kDa species and then to a more stable 25-kDa form, retains its IgE-binding activity [6].

[I 69071

* This work was supported with grants from the Medical Research Council (UK) and The Leukaemia Research Fund.

Correspondence: John Gordon, Department of Immunology, West Extension, The Medical School, Vincent Drive, Edgbaston, Birming- ham B15 2TJ, GB

Abbreviations: BCGF B cell growth factor ELISA: Enzyme-linked immunosorbent assay IL: Interleukin P(Bu),: Phorbol dibutyrate sol-CD23: Soluble CD23

Recently, it has been suggested that IL 4-promoted IgE pro- duction might be strictly dependent upon the CD23 molecule [7]. In addition, the ability of certain CD23 antibodies to promote DNA synthesis in cultures of activated B cells has been linked to the generation of soluble CD23 fragments (sol- CD23) which have been claimed to function as a B cell-derived B cell growth factor (B-BCGF) (8-101. IL4 has also been described as a “BCGF [ll] and synergizes strongly with an anti-CDw40 antibody for sustaining the cycle of already cycl- ing cells [12].

The precise nature of the interplay between autocrine and paracrine pathways for B lymphocytes remains obscure. Partly in an attempt to address this issue, we have exploited the observation that phorbol esters and calcium ionophores mimic the biochemical second messengers generated upon the phys- iological stimulation of B cells via antigen receptors [13]. Using the agents phorbol dibutyrate [P(BU)~] and ionomycin, which bind reversibly, we have been able to dissect critically the biochemical requirements for each separate component of the B cell mitogenic response and analyze, in detail, the con- tributions from IL 4 and CD23 to these pathways. The findings reported herein strengthen the notion of a special relationship between these two important regulatory molecules in control- ling B cell responses and provide new insights into their modes of action.

2 Materials and methods

2.1 B cell cultures

Highly purified, resting tonsillar B lymphocytes were prepared by negative selections and density separations on Percoll (Pharmacia, Uppsala, Sweden) gradients exactly as described

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1988 0014-2980/88/1010-1561$02.50/0

1562

in detail elsehwere [ 141. Cultures were performed at 37 “C in a moist 5% C 0 2 atmosphere by placing B cells in RPMI 1640 containing 10% fetal calf serum (FCS), antibiotics and 5 X lo-’ M 2-mercaptoethanol at cell numbers and for times indicated in the text. For pre-activations, cells were cultured for 24 h with P(Bu), at 1 ng/ml (Sigma, Poole, Dorset, GB) and ionomycin at 0.8 pglml (Calbiochem, La Jolla, CA) before removal of the stimulating agents by washing 4 x in RPMI 1640 containing 1% FCS followed by re-culture in medium with additions indicated in the text. IL4 was added as purified recombinant material, BCGF as a commercial prod- uct containing a cocktail of activities (Cellular Products, Buf- falo, NY) and IL2 as a recombinant product (Boehringer- Mannheim, Mannheim, FRG). The anti-CDw40 antibody, G28-5, was a purified IgGl preparation kindly provided by Dr. J. A. Ledbetter (Oncogen, Seattle, WA) while the CD23 anti- body, EBVCS4, was used as an IgM fraction prepared from supernatants of a hybridoma generously donated by Dr. B. Sugden (McArdle Institute, Madison, WI). Soluble fragments of CD23 were prepared as described in Sect. 2.2.

J. Gordon, J. A. Cairns, M. J. Millsum et al. Eur. J. Immunol. 1988.18: 1561-1565

0.5 pCi = 18.5 kBq of [3H]thymidine ([3H]dThd) contained in 50 pl of medium. At the termination of culture, the amount of radioactivity incorporated was determined as described previ- ously [ 141. Results represent means of triplicate determina- tions which never varied by more than 10% of each other. Quantitation of soLCD23 fragments in culture supernatants was peformed using a previously described enzyme-linked immunosorbent assay (ELISA) [5] which employs a typical sandwich approach with the CD23 antibodies EBVCS4 and EBVCSS which recognize different epitopes on the cleaved molecule. Surface expression of CD23 on B cells was quanti- tated by fluorescence-activated cell sorter (FACS, Becton Dickinson, Sunnyvale, CA) analysis using an equimolar cock- tail of the CD23 antibodies EBVCS5 and MHM6 as first layer and a fluorescein-isothiocyanate-conjugated sheep anti-mouse Ig antibody in the second layer. Histograms were generated from analyzing 50000 cells for each sample.

3 Results

3.1 Analysis of B cell progression factors 2.2 Purification of sol-CD23

SoLCD23 fragments were prepared essentially as described elsewhere [5] and as indicated in Scheme 1. The starting mate- rial for these experiments was fresh pooled supernatants taken from cultures of Epstein-Barr virus-transformed B lympho- cytes growing at logarithmic phase and kindly supplied by A. B. Rickinson, Department of Cancer Studies, Birmingham, GB. Preparations stored at 4°C were used within 3 days of purification for the biological assays indicated.

2.3 Measurement of B cell activation

DNA synthesis was determined at times indicated in the text by pulsing 10’ B cells cultured in 200 pl of medium in 96-well microculture plates (flat bottoms = 0.32 cm2) with

Purification of soluble CD23

1 2 5 -

I

L I

MHMG-column

relute ( 3 M KSCN)

Scheme 1. Preparation and gel analysis of sol-CD23. Soluble frag- ments of CD23 were purified by affinity chromatography on MHM6 coupled to Sepharose 4B (Pharmacia), as detailed elsewhere and represented in the Scheme. Freshly isolated material was analyzed under reducing conditions by sodium dodecyl sulfate-polyacrylamide (14%) gel electrophoresis followed by silver staining as shown in track “a”. Identical material, but stored at 4°C for 2 weeks, was reanalyzed as shown in track “B”. The positions of molecular weight markers are shown.

The strategy adopted for the pre-stimulation assay is repre- sented schematically in Fig. 1. The ability of the candidate progression factors to promote DNA synthesis in the pre-acti- vated cells removed from their initial stimuli is shown for the optimal concentration found for each agent. Optimal induc- tion of DNA synthesis, obtained by replating cells in P(Bu)~ and ionomycin, is shown for comparison. In the absence of any continued co-stimulus only IL 4, commercial BCGF and the CDw40 antibody (G28-5) showed appreciable progression activity in this assay. When applied jointly with P(Bu)2 how- ever, all factors, now including IL2 and EBVCS4 (a previ- ously identified stimulating CD23 antibody) [ 101 , provided respectable enhancement of progression to DNA synthesis at 46-56 h. Under these conditions, the actions of EBVCS4 were equal to those of commercial BCGF and IL2. It should be

cpm x104

- .a 1 r

1’ Stimulation 2’ Factor Addition e l ; n

0 24 46

P(Bu)p +Csl

Figure 1. Assay for B cell progression factors. Resting B cells were cultured at 5 X ldiml for 24 h with P(Bu)? at 1 ngiml and the calcium ionophore (CaI) ionomycin at 0.8 pgiml prior to washing in order to remove the activating signals. Washed cells were re-cultured with either BCGF (10% v:v), IL4 (1000 unitsiml), JL2 (100 unitsiml), G28-5 antibody (1 pg/ml) or EBVCS4 (20 pg/ml) with (I) or without (0) P(Bu), (1 ng/ml) for a further 32 h receiving [3H]dThd for the final 10 h. Results are expressed as cpm of radioactivity incorporated and are representative of four similar expeirments.

Eur. J. Immunol. 1988.18: 1561-1565 Interleukin 4 and soluble CD23 as progression factors for human B lymphocytes 1563

Table 1. Influence of ionomycin on progression activities

Progression [3H]dThd incorporation (cpm) following factor co-culture with’)

Exp. 1 Exp. 2 Control Ionomycin Control Ionomycin

Control 7 15Ob’ 7 069 3 326 2 849 G28-5 32 856 116487 17 345 63 198 EBVCS4 14016 15 449 5 732 6 539 IL 4 61 514 67404 44240 47417 BCGF 41 941 42 599 15 797 16988 IL 2 32 555 28912 8 372 6 240

a) As for Fig. 1 but ionomycin (0.8 pg/ml) added during second co-

b) Results represented as mean values of triplicate determinations. culture stage as indicated.

noted that when cultures were continued with P(BU)~, which by itself did not augment significantly further stimulations of pre-activated cells, the progression activity of IL 4 approached that achieved under optimal conditions of stimulation.

Results detailed in Table 1 show the influence of ionomycin on the progression activity of the different factors when it was retained over the secondary period of culture. It can been seen that only stimulations elicited by the CDw40 antibody were significantly enhanced by co-culture with ionomycin. In the case of G28-5, the calcium ionophore consistently synergized with this activity resulting in a level of DNA synthesis 3-4-fold greater than achieved with the antibody alone.

3.2 Progression activity of soLCD23

Sol-CD23 was purified as outlined in Scheme 1 and as detailed in Sect. 2.2. Using this protocol, the major product isolated migrated with a M, = 25 kDa. As noted previously [9], an invariable “contaminant” of affinity-purified sol-CD23 prepa- rations was a band migrating with M, = 14 kDa. This material did not react with CD23 antibodies in Western blotting (J. A. Cairns, unpublished observation) raising the possibility that it was a distinct molecule co-purifymg with sol-CD23. However, there is now good reason to believe that the 14-kDa band represents a further degradation product of the 25-kDa mate- rial. As seen in track “b” of Scheme 1, when the material originally isolated and characterized in track “a” was stored for 2 weeks at 4”C, all the 25-kDa product had now been converted to the 14-kDa species. Furthermore, identical pat- terns of banding have been observed for material isolated by affinity chromatography from culture supernatants of mouse L cells transfected with recombinant DNA encoding the CD23 gene (kindly provided by T. Kishimoto, Osaka, Japan). Therefore, we feel confident that the activity attributed to the affinity-purified preparations used in this study is contained within the soLCD23 molecule.

The acitivity of sol-CD23 in the pre-stimulation assay described above is shown in Fig. 2. In the absence of P(Bu)* at the second stage, soLCD23 augmented DNA synthesis to a modest level at concentrations above 2 ng/ml and reached a maximum above 50 ng/ml. In continuous co-culture with P(Bu)z after the prestimulation stage the progression activity of sol-CD23 was much more pronounced both reaching a

t \ 32 t i

t \

\ \ \ \

\ 6

300 60 12 2A 0.5 0

DGE Soluble CD23 (ng/ml)

Figure 2. Progression activity of sol-CD23. Cells were pre-stimulated for 24 h as in Fig. 1, washed and replated in soLCD23 with (m) and without (a) P(Bu), at 1 nglml to give a total incubation period of 56 h. Incorporation of [3H]dThd was assessed over the final 10 h as in Fig. 1.

higher plateau and requiring lower concentrations to do so. As found when using the stimulating CD23 antibody, a joint application with ionomycin failed to augment the progression activity of the soLCD23 preparation. It should be noted that, on degradation to the 14-kDa species, sol-CD23 appeared to lose all biological activity in these assays (results not detailed).

3.3 Dichotomy of IL 4 action on phorbol ester-driven stimulations

At doses higher than those used in synergistic combinations with ionomycin, tumor-promoting phorbol esters are directly mitogenic for human B lymphocytes [14]. Augmentation of this mitogenic effect has been used previously to monitor the co-stimulatory activity of CD23 and CDw40 antibodies [15]. In the present study, we found that, when added jointly at the commencement of culture with high doses of phorbol ester, soLCD23 similarly enhanced the subsequent DNA synthesis occurring (experiments not detailed). Unexpectedly, however, co-stimulation of resting B cells with IL4 and phorbol esters led to a consistent down-regulation of the response elicited by phorbol ester alone. The details of such an experiment are illustrated in Fig. 3. When added concomitantly with P(BU)~ at 10 ng/ml, IL 4 depressed the subsequent mitogenic response in a dose-dependent fashion with maximal inhibition being reached at 1000 unitslml of IL4. When the addition of IL4 to the P(Bu),driven stimulations was delayed by 24 h, it was now seen to enhance the DNA synthesis in the cultures. The dose-response curve for these later, positive effects essentially mirrored those of the earlier inhibitory actions (Fig. 3). It can also be seen that pre-culture of the resting B cells with IL4 24 h prior to the addition of P(BU)~ dampened the response even further than when added jointly so that the ensuing DNA synthesis was now approaching background levels.

1564 J. Gordon, J. A. Cairns, M. J. Millsum et aI.

- a, = A %

Eur. J . Irnmunol. 1988.18: 1561-1565

/ a A

I \ 0 'D-•

f03L- 0 125 250 500 1000 2000

IL4 [units/mO Figure 3. Influence of IL4 on phorbol ester stimulations. Resting B cells were cultured with 10 ng/ml of P(Bu), for 56 h receiving a [3H]dThd pulse for the final 10 h. The amount of radioactivity incor- porated is presented on a loglo scale. IL4 added simultaneously with P(Bu), (A); IL4 added 24 h prior to P(Bu), (0); IL4 added 24 h after P(B42 (.I.

3.4 Control of CD23 expression by IL 4 and phorbol ester

Both IL4 and phorbol esters have been shown to induce CD23 on resting B cells [4, 101. We therefore approached the possi- bility of whether the dichotomous effects of JL4 on phorbol ester stimulations noted above might be reflected at the level of CD23 expression and release possibly accounting for some of the observed effects.The results depicted in Fig. 4 confirm that IL4 and P(Bu)z are individually capable of promoting

n j/ml lih 35ng/ml

Urn' un \ 1 OSng/ml

linear fluorescence intensity

Figure 4. Analysis of surface CD23 expression and sol-CD23 produc- tion. Cells were cultured with IL4 (1000 unitshl) andor P(Bu), at 1 n g h l for 48 h as follows: (a) control cultures; (b) P(Bu), only (c) IL4 only; (d) IL4 followed by P(BU)~ at 24 h; (e) P(Bu), and IL4 together (f) P(Bu), followed by IL4 at 24 h. Histograms were gener- ated from 50000 cells, control staining is shown to the left, staining for surface CD23, to the right. The amount of sol-CD23 detected in the corresponding culture supernatants at 48 h is also indicated.

CD23 expression at the resting B cell surface. Furthermore, the level of induction did not alter greatly when the actions of IL 4 and P(BU)~ were combined nor when IL 4 was followed by P(BU)~ for 24 h later. However, when cells were pulsed first with P(BU)~ for 24 h then incubated with IL4, the level of CD23 expression rose dramatically (Fig. 4). This change in surface expression was accompanied by a large increase in the amount of sol-CD23 being released into the culture medium. Thus, the same sequence of additions which had been shown to generate high rate DNA synthesis also led to a substantial up-regulation of CD23 production and release.

4 Discussion

The findings described in this report firmly establish both T cell-derived IL 4 and B cell-derived sol-CD23 as progression factors for activated human B lymphocytes. They also indicate the possibility that a large part of this activity for the former might be reflected in its ability to promote high-rate produc- tion of the latter. Even if this were the case, it is clear that IL4 conveys information to the B cell additional to that attribut- able solely to its capacity for induction of sol-CD23. Thus, IL 4, but not sol-CD23, down-regulated phorbol ester-driven mitogenesis when added early to cultures while the pro- gression activity of sol-CD23, but not that of IL4, on pre- activated cells required a joint application with phorbol ester to reach significant levels. The dichotomy of IL4 effects on phorbol ester responses are of interest and confirm the notion that individual cytokines can invoke different biological responses within a given lineage depending upon the activa- tion status of the cell [I]. These observations might also pro- vide some clues as to the biochemical messengers through which activated IL 4 receptors convey their information to the inside of the cell.

A particularly dramatic finding of the study was the level of CD23 up-regulation induced by IL4 in B cells which had been pre-activated with phorbol ester for 24 h. The amounts found both on the surface of the cells and in the supernatants of the cultures- approached those associated with B lymphoblastoid cells generated by transformation with Epstein-Barr virus. It would seem unlikely that such large amounts of CD23 did not contribute to the progression activity of IL 4 under these con- ditions but we cannot say whether its production is an essential feature of the response. To answer this, a neutralizing CD23 antibody is required but, to date, none of those we have tested has been able to block the progression activity of the sol-CD23 molecule. All currently available antibodies have, however, been raised to membrane-bound CD23 and it is possible that the active site of the soluble molecule is generated only on cleavage and, therefore, would presumably be associated with a neo antigen. We are currently generating antibodies against sol-CD23 directly and hope to address this possibility in the near future.

The mechanism by which CD23, either in its membrane-bound or soluble form, delivers it growth-promoting activity is also unknown. One possibility is that, once cleaved, sol-CD23 re- associates with the cell surface via a specific receptor. It could be that such an association might exploit the potential lectin- like function of CD23 hinted at from its homology with other animal lectins [6]. Alternative mechanisms of stimulation do exist, however, and we are currently exploring the possibility that once cleaved, CD23 itself is endowed with proteolytic

Eur. J. Immunol. 1988.18: 1561-1565 Interleukin 4 and soluble CD23 as progression factors for human B lymphocytes 1565

activity which acts either on more membrane-bound CD23 or on other target molecules. In this regard, it is of interest that, even when highly purified, the 25-kDa form of CD23 degrades readily to lower molecular weight forms. The rapid loss of biological activity observed upon such degradation has made full functional characterization of soLCD23 problematical.

The requirements for the progression activity of soLCD23 appeared identical to those of a stimulating CD23 antibody. We have interpreted stimulation by appropriate CD23 anti- bodies as reflecting their ability to promote cleavage of the membrane-bound molecule. Although we have no reason as yet to discount this hypothesis, in light of the above comments it is worthwhile to consider an alternative mechanism in which certain CD23 antibodies protected active soLCD23 from auto- proteolysis to inactive forms. This would both promote the growth-stimulating activity and also given the appearance of more 25-kDa CD23 being generated as was observed in an earlier study [lo]. Again, such a possibility is being currently investigated.

Finally, it is of interest that only the signal delivered via O w 4 0 was augmented by the calcium ionophore ionomycin. The possibility that the Bp50 antigen described by CDw40 is associated with a kinase activity has been mooted [16] and this might account for such a stimulus co-operating with a Ca2+ signal. Similarly, the synergistic interaction of the CD23-deli- vered stimulations with phorbol esters suggest that they may invoke elevations in free Ca2+ levels. Interestingly, 12-kDa BCGF has recently been shown to promote such change in B lymphocytes [17] and CD23 has been linked to BCGF responses [ 181,

Received June 15, 1988.

5 References

1 Gordon, J. and Guy, G. R., Immunol. Today 1987. 8: 339.

2 Gordon, J., Ley, S. C., Melamed, M. D., English, L. S. and Hughes-Jones, N. C., Nature 1984. 310: 145.

3 Gordon, J., Ley, S . C., Melamed, M. D., Aman, P. and Hughes- Jones, N. C., J . Exp. Med. 1984. 159: 1554.

4 Defrance, T., Aubrey, J. P., Rousset, F., Vanbervliet, B., Bon- nefoy, J . y . , Arai, N., Takebe, y . , Yokata, T., Lee, F., Arai, K., De Vries, J . and Banchereau, J., J. Exp. Med. 1987. 165: 1459.

5 Cairns, J., Flores-Romo, L., Millsum, M., Guy, G., Gillis, S., Ledbetter, J. A. and Gordon, J . , Eur. J. Immunol. 1988.18: 349.

6 Yamasaki, K,, Kaisho, T., Uchibayashi, N., Hardy, R. R., Hirano, T., Tsunasawa, S. , Sakiyama, F., Suemura, M. and Kishimoto, T., Cell 1986. 47: 657.

7 Yokota, T., Arai, N., De Vries, J., Spits, H., Banchereau, J., Zlotnik, A, , Rennick, D., Howard, M., Takebe, Y., Miyatake, S., Lee, F. and Arai, K.-I., Immunol. Rev. 1988. 102: 137.

8 Gordon, J . , Rowe, M., Walker, L. and Guy, G., Eur. J . Immunol. 1986. 16: 1075.

9 Swendeman, S. and Thorley-Lawson, D. A., EMBO J . 1987. 6: 1637.

10 Guy, G. R. and Gordon, J . , Proc. Natl. Acad. Sci. USA 1987. 84: 6239.

11 Defrance, T., Vanbervliet, B., Aubry, J. P., Takebe, Y., Arai, N., Miyajima, A., Yokota, T., Lee, F., Arai, K., De Vries, J . and Banchereau, J., J . Immunol. 1987. 139: 1135.

12 Gordon, J., MiUsum, M. J., Guy, G. R. and Ledbetter, J . A., Eur. J. Immunol. 1987. 17: 1535.

13 Wang, F., Gregory, C. D., Rowe, M., Rickinson, A. B., Wang, D., Birkenbach, M., Kikutani, H., Kishimoto, T. and Kieff, E., Proc. Natt. Acad. Sci. USA 1987. 84: 3452.

14 Walker, L., Guy, G. R., Brown, G., Rowe, M., Milner, A. E. and Gordon, J . , Immunology 1986. 58: 583.

15 Gordon, J. , Webb, A. J., Walker, L., Guy, G. R., Paulie, S., Ehlin-Henriksson, B. and Rosen, A., in McMichael, A. J. et al. (Eds.), Leucocyte Typing III, Oxford University Press, Oxford 1987, p. 426.

16 Ledbetter, J . A., Rabinovitch, P. S. , June, C. H., Song, C. W., Clark, E.A. and Uckin, F. M., Proc. Natl. Acad. Sci. USA 1988. 85: 1897.

17 Einfeld, D. A., Brown, J . P., Valentine, M. A., Clark, E. A. and Ledbetter, J . A., EMBO J . 1988. 7: 711.

18 Gordon, J., Webb, A. J. , Walker, L., Guy, G. R. and Rowe, M., Eur. J. Immunol. 1986. 16: 1627.

1566 Eur. J. Immunol. 1988

All papers in this issue start on a new right-hand page, instead of immediately following the preceding article. Thus it may happen that part of an article's last page or even a whole (left-hand) page is blank. However, this does not result in higher costs or shorten the issue's contents. Rather, it allows faster publication, speeds up the production of reprints, and helps cutting cost increases.

I