PROLIFERATION OF RABBIT PERITONEAL HISTIOCYTES ASREVEALED BY A UTORADIOGRAPHY WITH TRITIA TED

THYMIDINE*

By MOSHE ARoNSONt ANI) SANFORD ELBERG

DEPARTMENT OF BACTERIOLOGY, IJNIVERSITY OF CALIFORNIA, BERKELEY

Communicated by K. F. Meyer, December 7, 1961

An effort has been made to increase our knowledge of the proliferating capacityof the rabbit peritoneal histiocyte which we have employed for several years in astudy of the infection process by species of Brucella and Mycobacterium.1-3 Con-flicting results and interpretations of data obtained from cell cultures concerningthe response of host cells in vitro to Brucella infection4' 5 revealed a need for attentionto the fate of the host cell.6 Results obtained from experiments employing "fly-ing coverslip" techniques differ in kind and degree from results obtained in micro-chambers such as the Mackaness cell. It therefore seemed to be of importance tolearn the extent of the histiocyte's ability to multiply under the conditions of"flying coverslip" technique, employing autoradiography with tritium-labeledthymidine to attempt to define the proliferative tendencies of the rabbit peritonealhistiocyte. A later report will compare in detail the behavior of histiocytes in bothkinds of cell maintenance conditions.

Several investigators7-9 have demonstrated the advantages to be gained from theuse of tritium for autoradiography. The incorporation of such a label into deoxy-ribonucleic acid has been useful in distinguishing production of new cells from pro-duction of new cellular constituents.

In light of the disagreement concerning multiplication of rabbit peritoneal histi-ocytes,6' 10 this cell was studied intensively because of its importance as a cell typein resistance and immunity, and because it has proved to be a very useful tool inthe study of cellular immunity in vitro.1-Methods.-A nimals: Rabbits, 4-5 lb, were injected intraperitoneally with 50 ml Klearol.

Inflammatory cells were harvested by the method of Fong et al.1Tritiated thymidine (0.5 microcurie/gram weight in 50 ml modified Tyrodet at 370C) was in-

jected intraperitoneally. The specific activity of the undiluted tracer was 0.36-2.7 curie/millimol.Smears of the cell suspension were fixed in absolute methanol for 15 min at 370C.The cells were distributed in Leighton tubes (0.25 X 106 cells per tube in 2 ml of 40% homologous

serum in Tyrode solution) and fixed after 3 hr. In this way one is provided with two sources ofmaterial. If the smears were too thick and the radiation was self-absorbed, the coverslip prepara-tion was available.

Autoradiography: The bottom part of slides was coated with gelatin; the coverslips weremounted (by means of "Permount") on gelatin-coated slides. AR 10 Kodak film was used inthe following way. In the darkroom the film was cut into strips which were floated on waterfor 3 min. The film was spread smoothly on the face of the slide. The covered slides were driedby a stream of air and then stored in a box in the refrigerator during the entire exposuretime, which varied from 5 to 180 days.

In the darkroom D19B developer was used for 5 min, followed by water for 30 sec. Fixationrequired 20 min in F5 acid fixer. Slides were rinsed with 10 changes of distilled water. Certainfilm required 15 min in 50% ethanol followed by 75% for 15 min for better staining. Slides weredried in air and stained with a modified Giemsa stain for 2 hr.§Results.-The relation between the time of injection of inflammatory stimulant and

the extent of incorporation of tracer material was studied for 5 days of the inflamma-

208

VOL. 48, 1962 MICROBIOLOGY: ARONSON AND ELBERG 209

tory response. At 5 days, 90 per cent of the cells were the histiocytic type, although theexudate continued to show a dynamic process 8, 10, and even 15 days after the oil in-oculation. Thus the results in Table 1 reflect the initial response plus the later stage

TABLE 1PER CENT LABELED CELLS UNDER DIFFERENT INJECTION SCHEDULES

Per cent of cellsVolume Per cent with label after

Expt. Days Histiocyte of cells with 180 days' ex-no. 0 1 2 3 4 5 yield Tyrode label posure of film

59S oil T(5h) H 1. 108 50 4870II oil T H 2.108 50 >40691I oil T H 4.107 50 851W oil T H 7.106 50 3568I oil T(4h)H 2.106 50 351T oil T(lh)H 7.106 50 241 oil T H 1.108 10 64710 oil T H 5.107 10 15 414750 oil T H 5.107 50 25 4356NB oil T H 1.1.108 50 6 -59T2 oil T H 2.108 50 1759W1 oil T H 1.5.108 50 459W2 oil T H 2.108 50 2.5 -62I oil T(4h)H 2.108 50 462II oil T(3h)H 2.107 50 3T = injection of tracer, H = harvesting, h = hours.

of a dynamic equilibrium between dying and newly arriving histiocytes (irrespectiveof point of origin). This point deserves emphasis because in most of the experi-ments the cells were harvested at a fixed time after the irritant injection (5 days)but at varying times after the tracer infection. It is likely that cell harvests fromanimals injected with tracer 3-5 days after the oil would show a higher degree oflabeling.When the tracer was injected 1-2 days after the oil and the cells harvested 3-4

days later, a much higher percentage of labeled cells was obtained than when theharvesting was carried out a few. hours after injection of the tracer.

In the short-term experiments, in which the cells were harvested within hoursafter the tracer injection, only a low percentage of label was obtained. Those cellswhich were labeled at all usually showed a high degree of labeling (hundreds ofgrains) and often turned out to be promonocytes. Generally where promonocytesappeared in the preparation, a high percentage of them was labeled.

It is also of interest in Table 1 that slides which had been exposed for 6 monthsshowed a much higher percentage of label. When the slides were scored separatelyfor low and high label (Table 2) the cells in experiment 4710 and 475° showed thatall of the "extra" grain either would not have appeared at all or would have beenconsidered too low after 10 days. Thus the data presented here are to be regardedas minimal values. It is also highly likely that "self-absorbency" began to operatein view of the thickness of the preparations.

Observations made earlier in Mackaness chambers1' 2 (that histiocytes multiplyin vitro) are in agreement with the results reported here, obtained by a method whichallowed one to determine whether most of the cells were multiplying or only asmall minority. Since the in vivo studies showed that a very active multiplicationoccurred during the first 2 days of the response, one may predict that histiocytes

210 MICROBIOLOGY: ARONSON AND ELBERG PROC. N. A. S.

TABLE 2GRAIN DISTRIBUTION OF LABELED HISTIOCYTES

Experiment Percentage ----number 5-20 20-50 50-200 20059S 1007011 1006911 66 10 4 2051W 55 35 10681 0 25 50 2551T 0 10 20 7041 34 24 35 74010 90 7 3 0475° 80 3 12 456MB 80 2 0 1859T2 26 8 26 4059W1 6 4 16 7459W2 6 26 56 1262I 6 6 6 826211 0 10 23 67

harvested at this stage would start to incorporate tritiated thymidine in vitrosooner than those harvested after 5 days.

In order to test this prediction, 2- and 5-day-old histiocytes were harvested andtreated conventionally. In the case of cell suspensions obtained in 2 days, therewere many polymorphonuclears, lymphocytes, and promonocytes. The numberwas adjusted according to histiocytes and promonocytes. The cells were thenincubated up to 4-5 days in the presence of 0.1-1 microcurie thymidine/ml medium.For autoradiography the slides were exposed for 5-10 days.

(a) Five-day histiocytes: The-results are shown in Table 3, where the data on500 cells were gathered. Each result is the mean of 2 coverslips prepared in dupli-cate.

TABLE 3PER CENT OF LABELED HISTIOCYTES CULTURED FROM SUSPENSIONS HARVESTED AT VARIOUS

TIMES FROM RABBITHours - Five days - Two days - Lung histiocytes--

in Expt. Expt. Expt. Expt. Expt. Expt. Expt. Expt. Expt. Expt.culture 152 160 166 167 158 161a 161b 50 152 45

3 2 1 0 0 3 3 5 5 1 024 1 1 3 1/2 20 4 8 4 2 2748 26 5 50 37 53 5 4572 54 16 71 66 78 1596 44 - 80 80 14

The results agree at the early readings; i.e., very little incorporation at 3 and 24hr and a considerable increase at 48, but they differ considerably in the older cul-tures. Apparently, optimal conditions of culture in Leighton tubes had not beenachieved for rabbit peritoneal histiocytes.

(b) Two-day histiocytes: Smears were made from the preparations harvested 2days after the oil injection. These smears, stained by Wright's method, revealedmany large cells, more basophilic, which have nucleoli (Fig. 1). They were pre-sumed to be promonocytes. When such preparations were incubated invitro in the presence of tritiated thymidine, there was already a considerable numberof labeled cells at 24 hrs. Many of the labeled cells were larger than 5-day-old cells,did not look like fibroblasts, and were judged to be promonocytes (Fig. 2). It is ofinterest to note that lung histiocytes from tracheal washings harvested 1, 2, 3,4, and

VOL. 48, 1962 MICROBIOLOGY: ARONSON A.ND ELBERG 211

.:.:;:;;.........;.AM&~ ~ ~FIG. 1Large basophilic cells harvested two days after injection of oil.

FIG. 2.-Labeled cells present in exudate two days after oil injection.

5 days after injection of the tracer failed to reveal evidence of labeling. Further-more, no labeled cells were seen in blood smears prepared 4 hr or 4 days after the in-jection of the tracer.

In the course of this work we modified the fixation method of histiocytes cultivatedin Leighton tubes, to prevent the majority of the cells from slipping off the cover-

212 MICROBIOLOGY: ARONSON AND ELBERG PROC. N. A. S.

slip when it is heated in air at 370C. Methanol worked well, but the cells shrunk insize and were thus rendered useless for Machiavello staining. Our technique em-ployed washing at 370C and flattening the cells by applying a warm agar-block,followed by fixation with warm methanol (370C).

0V

*:

FIG. 3.-Cytoplasmic bridges and evidence of intercellular exchange.

Fcbde ce x

FIG.4.-Ctoplsmicbrides ad evdenc of inte.~l xhne

VOL. 48, 1962 MICROBIOLOGY: ARONSON AND ELBERG 213

Histiocytes so fixed presented some new features. The pseudopodia were clearlyvisible; very often, two neighboring cells seemed to be joined through them by"cytoplasmic bridges" (Fig. 3). When parasitized histiocytes were used, the bac--teria could be seen in the "bridges." Figure 4 shows a cytoplasmic bridge betweena histiocyte and a "giant cell." The histiocytes in this case had been infected withListeria monocytogenes. This observation, provided it is not due to an artifact,raises an interesting possibility: that infected histiocytes "transmit" intracellularparasites and perhaps antigenic material in general into other histiocytes.Discussion.-The earlier statement that peritoneal macrophages are nonmulti-

plying cells6 has been found to be correct up to 24 hr but not for longer periods ofstudy. There was no incorporation of label at 24 hr, but at 48 hr a considerablenumber of the cells was labeled. Since DNA synthesis precedes cell division, oneshould expect to find labeled cells somewhat before a cell increase is noted in culturechambers, which usually occurs between 48-72 hr. Additional evidence for in vitromultiplication came from studies on cells which were labeled in vivo and maintainedin a label-free medium in vitro. After 48 hr in culture, there were many binucleatecells which contained an equal amount of label in both nuclei. In these prepara-tions one could occasionally see mitotic figures and dividing cells.These results also show that practically all of the histiocytes were capable of cell

division when properly maintained. This excludes the possibility that the increaseof cells observed earlier in the Mackaness chamber2' I was due to "contamination"with other cells which multiply very rapidly. Unfortunately, the data suggestthat not all the factors required for good growth in Leighton tubes are yet undercontrol, with resulting discrepancies between experiments.The experiments with cells harvested 2 days after the oil stimulus suggested

that the abundant promonocytes, which were the first cells showing an in vivouptake of label, were also the first to incorporate it under the in vitro con-ditions.Summary.-Tritium-labeled thymidine has been used to label nuclei of rabbit

peritoneal histiocytes in vivo and in vitro. After a lag period of 48 hours, cells main-tained in vitro started to incorporate the tracers, and ultimately most of them werelabeled. Injection of the tracer into the peritoneal cavity after the injection of anirritant resulted in the appearance of a high percentage of labeled cells; suggestingproliferation in the cavity.

We are greatly indebted to Norbert Schachter for technical assistance, and to John Shutes,Naval Biological Laboratory, for photographic assistance.

* This research was supported by the National Science Foundation.t Member of the Israeli Institute for Biological Research, Ness-Ziona.t NaCl, 7.7 gm; KCl, 0.2 gm; MgSO4, 0.1 gm; dextrose, 4 gm; NaH2PO4, 0.05 gm; water,

750 ml. To this is added: NaHCO3, 0.5 gm; water, 250 ml which has been separately sterilizedby filtration.

§ Communicated to us by Professor T. Crocker. Details on request.1 Fong, J., P. Schneider, and S. Elberg, J. Exp. Med., 104, 455 (1956).2 Elberg, S. P. Schneider, and J. Fong, ibid., 106, 545 (1957).3 Fong, J., D. Chin, H. J. Akiyama, and S. Elberg, ibid., 109, 523 (1959).4Freeman, B., D. J. Kross, and R. Circo, J. Infect. Dis., 108, 333 (1961).6 Jenkins, C., and B. Benacerraf, J. Exp. Med., 112, 403 (1960).6 Watts, J. H., and H. Harris, Biochem. J., 72, 147 (1959).

214 MICROBIOLOGY: W. F. GOEBEL PROC. N. A. S.

7 Hughes, W., V. P. Bond, G. Brecher, E. P. Cronkite, R. B. Painter, H. Quastler, and F. G.Sherman, these PROCEEDINGS, 44, 476 (1958).

8 Taylor, J. H., P. S. Woods, and W. Hughes, ibid., 43, 122 (1957).9 Speirs, R., E. Speirs, and V. Jansen, Proc. Soc. Exp. Biol. Med., 106, 248 (1961).10 Osgood, E., and C. Ashworth, Atlas of Hematology (San Francisco: J. W. Stacey, Inc., 1937).

THE CHROMATOGRAPHIC FRACTIONATION OF COLICINE K*BY WALTHER F. GOEBEL

THE ROCKEFELLER INSTITUTE

Communicated December 21, 1961

The elucidation of the nature of colicines has been an engrossing problem whichthis laboratory first undertook some ten years ago.' Colicines are potent bacteri-cides which are elaborated by many strains of Enterobacteriaceae.2 Our work hasrevealed that at least two of the colicines, V and K, either are intimately associatedwith the type specific 0 antigen of the bacilli from which they are derived or areidentical.3' 4 It will be recalled that 0 antigens are complex macro-molecules con-stituted from protein, lipid, and polysaccharide.6 It is these substances whichendow many Gram negative bacilli with immunological specificity and toxic prop-erties as well. Throughout our research, we have never lost sight of the possibilitythat the colicine itself might indeed be a separate chemical entity which merelyaccompanies the 0 antigen. It is our belief, however, that this is not the case,largely because all efforts to separate the two have been fruitless.The use of substituted cellulose derivatives for the separation of complex mix-

tures of biological origin has proved so successful in the hands of others,6 that wehave employed them in an attempt to separate colicine K activity from the 0antigen of the colicinogenic bacillus E. coli K235. The following is an account ofour investigation.

Preliminary experiments revealed that antibacterial activity could be readilyadsorbed by DEAE cellulose from a solution of colicine K in tris buffer 0.05 Mat pH 8.0 (ratio of adsorbent to colicine 50:1). The colicine activity was releasedby elution with the same buffer at lower pH and at higher ionic strength (0.3 Mtris-0.5 M NaCl, pH 7.2.). A large scale chromatographic experiment was nowperformed. 60 gm of DEAE cellulosef free of UV-absorbing material, was sus-pended in 0.05M tris at pH 8.0. The slurry was packed in a column (4.5 X 40 cm)which was maintained at 40C and the cellulose washed with 2 liters of the samebuffer. A solution of purified colicine K (1.2 gm in 80 ml of 0.05 M tris, pH 8.0)was permitted to flow into the column by gravity. The collection of 100 ml samplesof effluent was started at this point (flow rate 70 ml/hr); the initial eluant was600 ml of the pH 8.0 buffer. The adsorbed material was eluted from the columnwith 7 liters of tris-NaCl buffer. The concentration of the latter was graduallychanged by allowing a solution of 0.3 M tris-0.5 M NaCl at pH 7.2 to flow intoan Erlenmeyer mixing flask 7 containing 4 1 of 0.05 M tris at pH 8.0. In all, 70