Znt. J. Cancer: 20, 353-358 (1977)

SPONTANEOUS HUMAN LYMPHOCYTE-MEDIATED CYTOTOXICITY AGAINST TUMOR TARGET CELLS. 11. IS THE COMPLEMENT RECEPTOR NECESSARILY PRESENT ON THE KILLER CELLS? Hugh F. PROS1, 2, Malcolm G. BAINESa, and Mikael JONDAL Departments of Radiation Oncology and Pathology; Correspondenceshould be addressed to Dr. Hugh Pross, Ontario Cancer Foundation, Kingston Clinic, Kingston, Ontario, Canada K7L 2 V7; Department of Obstetrics and Gynecology, Queen's University, Kingston General Hospital, Kingston, Ontario, Canada; and Department of Tumor Biology, Karolinska Instituie, Stockholm, Sweden

The effect o f various lymphocyte depletion techniques on SLMC effector cell activity against the cell line K562 could be attributed t o the effects that these procedures had on the proportions of FclC3 receptor-bearing cells in the preparations. Depletion of cytotoxic activity by removal of Fc receptor-bearing cells could be augmented by the presence of complement on the 'IS-EA indicator cells used for rosette depletion. Two methods o f assessing t h e presence of surface receptors on cytotoxic cells are presented: (1) linear regression analysis o f spontaneous cytotoxicity plotted as a function of the proportions o f receptor-bearing cells remaining after various depletion techniques; and (2) the use of a mathematical formula t o estimate the proportion o f cytotoxic cells w i t h any particular receptor, based on the effects of receptorbearing cell depletion. Linear regression analysis of cytotoxicity vs the proportions of SRBC, Fc o r C3 receptor-bearing cells demonstrated that all cytotoxic cells had the Fc receptor, but that cytotoxicity could s t i l l occur in the absence o f cells w i t h the C3 receptor. This indicated that t h e C3 recept o r i s n o t necessary f o r the demonstration of SLMC activity. The mathematical formula t o predict cytotoxicity after depletion of C3 receptorbearing lymphocytes showed, however, that a significant proportion o f the Fc receptor-positive killer cells were also C3 receptor-positive. No direct evidence was found t o support a role for T cells in SLMC, although E rosette depletion resulted in lower than expected cytotoxicity i n view of the number o f Fc receptor-bearing cells remaining in the preparation.

In a recent publication, West et al. (1977) have gone to some lengths t o evaluate our contention that the human SLMC killer cell is complement receptor-, Fc receptor-positive (Pross and Jondal, 1975a, b; Jondal and Pross, 1975). To make comparisons more meaningful, they have used a modification of the assay system which we use routinely, involving the myeloid cell line, K562 (Lozzio and Lozzio, 1973), as the target cell in a 61Chromium release assay. In their work they have confirmed two important points: first that this assay system is highly sensitive in terms of detecting human SLMC activity, and second that the killer cell has the Fc receptor. The presence of the Fc receptor

on these cells has been reported by several groups (reviewed in Pross and Baines, 1976), although its role in SLMC remains to be established. The point at issue is whether or not the complement receptor is also necessarily present on these cells (de Vries et al., 1974; Jondal and Pross, 1975; Peter et al., 1975). Kiuchi and Takasugi (1976) have indicated that complement receptors may not be present (based on experiments in which they failed to recover cytotoxic cells from 19s-EAC rosette enriched cell pellets), and West et al. (1977) and Hersey et al. (1975) have drawn similar conclusions using several different method? of complement-receptor-bearing cell depletion. The point is an important one since cellular receptors can be used to remove SLMC killer cells from other cytotoxic systems (Jondal and Pross, 1975; Svedmyr and Jondal, 1975). The present paper attempts to deal with this apparent discrepancy between ourselves and these investigators. Data are presented based on rosette depletion techniques designed to remove different types of receptor-bearing cells, and using new sources of 19s rabbit anti-SRBC. With these sera, 19s-EAC rosette-forming cell depletion was found to result in only partial removal of SLMC activity, complete removal being made possible by depletion of rosettes formed with SRBC coated with 7 s rabbit anti-SRBC or 7 s rabbit anti-SRBC plus complement. The method of analysis used in this brief communication represents a different approach to the study of surface receptors

Received: May 3, 1977 and in revised form June 27, 1977.

Abbreviations : SLMC: spontaneous lymphocyte mediated cytotoxicity against tumour targets found in normal human blood; T: thymus derived; SRBC, E: sheep erythrocytes; 7s-, 19s-EA: E coated with 7 s or 19s rabbit anti-E; EAC: EA coated with mouse complement; RFC: rosette-forming cell; Fc receptor: receptor for the Fc fragment of IgG; C3 receptor: receptor for the C3 component of complement; SE: standard error.

354

PROSS ET AL.

on cytotoxic cells, and indicates that a significant proportion of Fc-positive killer cells possess the C3 receptor. MATERIAL A N D METHODS

Cytotoxicity assay for human SLMC

19s anti-SRBC obtained from Cordis Labs., Miami, Fla., (19S-EA, Fig. 2, H.F.P., and M.G.B.). 7s-EAC or 19s-EAC were prepared by the incubation of 2 % EA cells with an equal volume of fresh undiluted mouse serum as a complement source for 20min at 37" C. The use of diluted serum as a complement source was found to result in rosettes with less red cells per rosette, adequate for rosette enumeration, but possibly less than adequate for depletion experiments.

Purified peripheral blood lymphocytes were obtained from heparinized blood by Ficoll-Isopaque gradient centrifugation (Boyum, 1968), followed by carbonyl iron plus magnetism treatment to remove adherent and phagocytic cells (Lundgren Statistics et al., 1968). The cells were tested in a 51Chr~mium Regression analysis of the correlation between release assay using K562 as the target cell, as previously described (Pross and Jondal, 19756; the frequency of various cell types and cytotoxicity was performed by the least squares method. The Jondal and Pross, 1975). significance of the regression was determined by Rosette-forming cell depletion techniques computing the F statistic from the ratio of the Rosette depletion was done by Ficoll-Isopaque explained mean square over the unexplained mean gradient centrifugation of lymphocytes subjected square (Sokal and Rohlf, 1969). to various rosetting techniques similar to those used for cell identification, but using larger numbers of lymphocytes and indicator cells (Jondal, 1974). A single depletion step could be expected to result in removal of 80-90% of the rosetting cell type. As a rule of thumb, the lymphocyte yield at the medium-Ficoll-Isopaque interface is approximately one-half to one-third of that expected after subtraction of the percentage of rosettes in the original preparation.

RESULTS

The role of complement in augmenting rosette depletion

Figure 1 illustrates an experiment designed to assess the relative efficiency of rosette depletion

Rosette formation

E rosettes, T cells, were formed by the method of Jondal et al. (1972), except that incubation proceeded for 15 min at 37" C followed by 60 min at 4" C. The rosettes were suspended very gently prior to counting or placing on the gradient for separation. 7s-EA and 19s-EAC rosettes (Fc and C3 receptor-bearing cells respectively) were formed by centrifugation of lymphocytes mixed with the appropriately treated SRBC indicator cells, followed by immediateresuspension (Fig. 2) or by resuspension after incubation at 37" C for 30 min (Fig. 1). Controls consisted of rosettes (if any) formed using similar techniques with uncoated SRBC (rapid E rosettes), or SRBC coated with 19Srabbit anti-SRBC(19SEA). EA, EAC and control preparations were resuspended more vigorously than E rosettes. Antibody and complement

For the preparation of rosette indicator cells a subagglutinating concentration of the following antisera against SRBC was used: hyperimmune rabbit anti-SRBC (7SEA); rabbit 19s anti-SRBC obtained 5 days after intravenous injection of 1 ml of 10% SRBC and purified by Sephadex G-200 chromatography (19SEA, Fig. 1, M.J.), and rabbit

4 2 1 : ]h1 ,1 , 2 c

a

,

I

,

Y

a"

20/1 10/1 511

20/1 10/1 5/1

20/1 1011 511

LymphocyteITorget Cell Ratio FIGURE1

The effects of EA and EAC RFC depletion on SLMC activity. The antibody-sensitized erythrocytes (19s-EA or 7s-EA) used in the depletions are indicated above each column. The symbols indicate the SLMC activity after depletion of RFC formed with EA cells prepared either with no complement (0), with heat-inactivated mouse complement ( x) (56" C , 30 min), or fresh mouse complement (0).The open triangles (A) indicate the cytotoxic activity of undepleted control lymphocytes subjected to identical centrifugation on Ficoll-Isopaque. These experiments were performed at the Karolinska Institute using the reagents described in Material and Methods". The standard error of the mean percentage chromium release did not exceed f 2 % .

355

COMPLEMENT RECEPTORS A N D SLMC KILLER CELLS

using 7S-EA, or 19S-EA, plus or minus complement, as a method of removing spontaneous cytotoxicity against K562. 19s-EA rosette depletion had no effect on cytotoxicity. 7S-EA rosette depletion was more effective than depletion of 19s-EAC rosettes, but with this antiserum, maximum depletion occurred with 7s-EAC rosettes. This study indicates that Fc receptor-bearing cell depletion is more effective than depletion which depends solely on the presence of the complement receptor on the killer cell. The binding of the killer cells to both 7s-EA and complement resulted in the most efficient removal of these cells after gradient centrifugation.

Lymphocyte subpopulation depletion and cytotoxicity Figure 2 illustrates one of four experiments in which a somewhat different approach was used to examine the relative importance of the various possible receptors on the SLMC cell. In this experiment, preparations were made of 7S-EA, 19S-EA, 19s-EAC (Cordis 19S), and E rosettes. An additional control preparation, called rapid E rosettes (to distinguish from conventional E rosettes), was made using 1 % SRBC in medium, with rosette formation and immediate resuspension being done as for the 7S-EA, 19S-EA, and 19s-EAC rosettes. It should be emphasized that all of the depletions were

I

40

A

404

35-

35

30-

30

25-

25

20-

20

m

0 Ln

-w (L

-iE e L *

+

15-

" c w

15

; 10-

a

5-

10

'\,

J

B ,

5 1 2 5/1 Lymphocyte/Torget Cell Rotio

lO/l

0

,

, A ,

, , 7

,

I

10 20 30 40 50 60 70 80 90 100

Percent R F C

FIGURE 2 One of four experiments designed to demonstrate the effect of various RFC depletions on SLMC. A: the control preparation, x, from a single donor, was depleted of the following RFC types: 0 , rapid E; I, E; A, 19s-EA; A , 19s-EAC; 0 , 7s-EA. The percentage 51Crrelease caused by the different rosette-depletedpreparations is shown at three lymphocyte: target cell ratios. B: linear regression analysis of various RFC proportions remaining in the preparations shown in Figure 2~ YS the cytotoxicity of the preparation at a lymphocyte: target cell ratio of 1O:l. The same total number of cells was present in each assay well. The calculated regression lines are indicated for the relationship between percentage RFC and SLMC activity as follows: 19s-EAC RFC (A-A); and E RFC (I-.). 7s-EA RFC (0-0); The cell recovery after rosette depletion and three washes was as follows (expressed as percentage of cells placed on the gradient): E RFC-depleted 19 %; 7s-EA-depleted 33 %; 19s-EAC-depleted 24%; 19s-EA-depleted 17 %; rapid E-RFC-depleted 18 %. The latter yields were lower than expected because some E-RFC depletion occurred in these three preparations, as shown in , horizontal group of three symbols indicates the Figure ZB.In Figure 2 ~ each proportions of 7s-EA (o), 19s-EAC (A), and E (I) RFC remaining in the preparations after depletion of the following rosette types (reading from top to bottom in Figure 2~)--19s-EA, rapid E, E, undepleted control, IgS-EAC, and 7s EA. These experiments were performed at Queen's University using the reagents described in " Material and Methods ". The standard error of the mean percentage chromium release did not exceed *2%.

356

PROSS ET AL.

performed on lymphocytes from the same donor, and at the same time, hence eliminating both donor and day-to-day variation in the assay results. The effects of these depletions on cytotoxicity are shown in Figure 2A, in comparison with control, unfractionated lymphocytes. In these experiments there was less reduction of cytotoxicity after 19s-EAC depletion, and a greater reduction of cytotoxicity after 7s-EA depletion than shown in Figure 1 . This is probably attributable to differences in the antisera used (see “ Material and Methods ”). Three effects on cytotoxicity are apparent from Figure 2~:enhancement(by E, 19s-EA, and rapid E rosette depletion), moderate reduction (by 19s-EAC depletion) and profound reduction (by 7s-EA depletion). Some of these effects, especially those involving enhancement by “ control ” preparations, are difficult to explain in terms of the anticipated effects of rosette depletion, until one examines the relative proportions of the different cell types remaining in the preparations after fractionation. To evaluate the re1ative.importance of these different cell types in mediating cytotoxicity, the percentage cytotoxicity toward K562 (1ymphocyte:target cell ratio of 1O:l) was plotted against the percentage rosettes in each depleted or control preparation (Fig. 28). Linear regression analysis was performed on the data and the correlation coefficients were calculated. The one aberrant point in each of the sets of data for 7s-EA and 19s-EAC was omitted from these calculations (see below). It can be seen in Figure 28 that there is a negative correlation between the proportion of E rosettes and cytotoxicity (r = -0.8432, p