- Email: [email protected]
J Mice
Immunobiol., vol. 199, pp. 23-38 (1998)
Department of Anatomy and Cell Biology, McGill University, Montreal, Quebec, Canada
Strain Differences in Natural Killer Cell-mediated Immunity among Mice: A Possible Mechanism for the Low Natural Killer Cell Activity of AIJ Mice ANDREA L. WHYTE and SANDRA
C. MILLER
Received July 7, 1997 . Accepted in revised form November 20, 1997
Abstract Natural killer (NK) cells are well established as fundamental elements in the early eradication of aberrant cells potentially leading to neoplasia. Moreover, it has also long been known that inbred strains of laboratory mice, as well as human individuals, demonstrate a wide range of NK cellmediated immune response even to the same tumor. In the present study, various parameters which could lead ultimately to high, or low, NK cell-mediated functional activity have been assessed. Mice of the A/] strain demonstrate very low NK cell tumor-lytic activity and correspondingly high incidence of lymphoma. By contrast, C57B1I6 mice demonstrate relatively high NK cell activity and virtually never develop lymphomas. The results of this study have revealed that the absolute numbers of splenic NK cells were significantly lower in A/] vs C57BI6/mice. Furthermore, the blood of A/] mice contained significantly fewer (30%) NK cells than did that of C57B1I6 mice. However, no significant difference between the 2 strains was found in the numbers of lymphocytic cells from NK cell-enriched fractions from the spleens, which possessed either the homing receptor MEL-H, or the integrin Mac-I, both essential surface molecules for transendothelial migration of lymphocytic cells from the circulation into organ parenchyma. Moreover, NK cells from both strains responded similarly to the NK cell stimulants, ATRA, indomethacin and interleukin-2. Finally, there was no significant difference between the 2 strains, in the numbers of lymphocytic cells in the bone marrow (including NK cells), which were radiolabelled with the DNA synthetic precursor, 3H-thymidine, indicative, thus, of equivalent levels of lymphocyte production by the bone marrow in the 2 strains. The observations collectively suggest that the low peripheral (spleen, blood) levels of NK cell-mediated functibnal activity found in the A/] strain of mouse at least, reflects either post-production, large-scale NK cell abortion/death, or a bone marrow-based microenvironmental deficiency which inhibits NK cells' exit from the bone marrow birth site.
Abbreviations: ATRA = All trans retinoic acid, rIl-2 - recombinant interleukin-2; ASGM-l'= Asialo-GM-1; DAB = diaminobenzidine, CI = cytotoxicity index; NK = natural killer; FCS = fetal calf serum. C1998
by Gustav Fischer Verlag
24 . A. L. WHYTE and S. C. MILLER
Introduction Natural killer (NK) cells have been characterized morphologically as granular, non-blast lymphocytic cells, small to medium in size, in all species studied (1-3). NK cells have the ability to lyse a wide variety of abnormal cells, i.e., tumor and virus-infected cells, and in vivo, there is little doubt that NK cells provide the primary defence or immunosurveillance against the spontaneous development of tumors (4-8). The origin of NK cells is the bone marrow (9-12). Newly-formed, mature NK cells are not themselves in cell cycle and leave the bone marrow birth site, randomly after their production (13). NK cells have an intramyeloid T t/2 of 8 h, followed by a circulation (blood transit) time of approximately 10 h while they are en route to the spleen wherein the T t/2 of NK cells is 24.5 h (13). Longevity, memory and recirculation are not characteristics of NK cells (13-15). The integrin, Mac-1 (CDllb/CD18), found on NK cells has been implicated, at least in vitro, for transendothelial migration, where it complexes with the endothelial ligand, ICAM-1 (16-20). The homing receptor gp90MEL-14 (Lselectin) permits loose, initial binding of lymphocytes with endothelium (21), and the receptor has recently been found on NK cells in vivo (22), suggesting a physiological role on NK cells, for this surface molecule long known to be essential for in vivo transendothelial migration of T and B lymphocytes. The homing receptor is soon shed while other, more firmly adhering, adhesion molecules such as Mac-1 are up-regulated to permit a tight lymphocyte-endothelial cell union immediately prior to transendothelial migration (23). There are 3 known potent stimulators of NK cell functional activity: the cytokine, interleukin-2, the drug, indomethacin and the metabolite of vitamin A, all-trans retinoic acid (ATRA). We have shown previously in mice which normally express relatively high levels of NK cell-mediated functional activity, that indomethacin and interleukin-2 augment the absolute numberslproduction of NK cells, leading, thus, to elevated functional levels of this cell in any organ tested (24-26). ATRA has been shown to enhance NK cell functional activity in mice and humans by up-regulating TNF-
Strain differences in NK cell immunity . 25
Collectively, the results have shown that, indeed, mature NK cells in AI] mice, are significantly lower in absolute numbers than in C57Bl/6 mice. Moreover, the NK cell numbers in both C57Bl/6 and AI] mice could be enhanced in vivo by interleukin-2, indomethacin but not by ATRA. However, lymphocyte production (NK and B cells) by the bone marrow (assessed by the frequency of lymphoid cells bearing a radiolabelled nucleotide precursor of DNA-s phase) was similar in both strains contrasting, thus, with the significantly lower levels of NK cells in the blood and spleen of AI] mice, and also with the significantly lower levels of all lymphoid cells found in the spleens of AI] relative to C57B1I6 mice. In AI] mice, thus, there appears to be a significantly lower emigration of newly generated NK cells (relative to C57B1I6 mice) from the bone marrow.
Materials and Methods Mice
A/J and C57B1I6 male mice (6-9 wk) were obtained from the Jackson Laboratories, Bar Harbor, ME, USA. Mice were housed in microisolator cages, maintained on standard complete mouse chow diet and remained undisturbed for 1 week prior to commencement of any experimentation. Tumor cell line
YAC-l, a lymphoma originating from A/Sn mouse strain, maintained in suspension culture consisting of RPMI 1640 medium with 10% FCS and 1% antibiotics and fungizone (GIBCO, Inc., Canada), served as the target of NK cells. Drug administration
Indomethacin (Sigma Chemical Co., St. Louis, MO, USA) was administered to mice through the drinking water. Five mg of the drug were dissolved in 1 ml absolute ethanol and diluted with water to give a concentration of 10 }lg/ml. Fresh solution was provided in drinking water every second day. No difference in the amount of water consumed between groups of mice given indomethacin, the indomethacin vehicle (ethanol in water), or untreated drinking water was observed as determined by measurements taken at 2 day intervals over the course of 1 wk. Cytokine administration
Human recombinant interleukin-2 (rl1-2), a generous gift from Hoffmann-LaRoche, Nutley, NJ, USA, was given as bolus intraperitoneal injections 3 times a day at 8 h intervals for 3 days. Recombinant 11-2, provided in the lyophilized state had a specific activity of 14xl06 Units/mg protein. A dose of 12xl03 Units in 0.1 rnl sterile phosphate-buffered saline, containing 1% homologous serum was administered at each injection. Toxicity levels, optimal dose-response curves, and injection frequency of rl1-2 in mice have been all well established by ourselves and others (24, 26, 40-43). Isotope administration
Tritiated thymidine, a radiolabelled precursor of DNA, readily incorporated by cells undergoing DNA synthesis, was given to mice intraperitoneally as a single dose of 1 }lCi/mg body weight, specific activity 20 Ci/rnM (ICN Biomedicals, Inc. CA, USA) in 0.5 rnl phosphate buffered
26 . A. L. WHYTE and S. C. MILLER saline. One h or 5 h later, mice were killed and cytospot samples of the spleen and bone marrow of 6-8 mice were processed for radioautography by the standard methods used in our laboratory (13, 44, 45). Administration of all-trans retinoic acid (ATRA)
ATRA (Sigma Chemical Co., St. Louis, MO, USA) was dissolved in commercially available corn oil (0.125 mg/0.1 ml), as the vehicle for oral administration immediately prior to administration, employing standard techniques of ATRA gavaging (46, 47). Five mg/kg body weight of ATRA was given every 24 h for 5 days. This time course was selected as more than adequate to ensure that any effect the agent might have on NK cell numbers/production would be readily evidenced given the short life span and rapid renewal of NK cells (13-15). In all cases, control mice were fed the vehicle only. Identification of homing receptor and integrin molecules on NK cells
Rat anti-mouse CDllb/CD18 (Mac-1) and rat anti-mouse L-selectin (MEL-14), purchased from Cedarlane Laboratories, ON, Canada), were diluted to optimal concentrations (based on dpse-response curves carried out in pilot experiments) of 1:5 and 1:30, respectively, prior to use. These primary antibodies were linked subsequently to biotinylated goat anti-rat IgG (Cedarlane Laboratories, ON, Canada) at concentrations of 1:100. Cells so stained were processed by the Vectastain method (Vector Laboratories, CA, USA), for identification of lymphocytes bearing either the homing receptor or the integrin. Tissue sampling
Spleens from experimental and control mice were removed and pressed through stainless steel screens into Petri dishes containing 5-10 ml ice-cold RPMI containing 10% FCS. From the same mice, the femurs were removed and flushed repeatedly with fresh RPMI + 10% FCS. Both spleen and bone marrow cell preparations were layered over 1 ml NCS for 5 min to allow sedimentation of large debris through the NCS. The cell suspensions were then removed, centrifuged for 5 min (500xg, 4 0c) to yield a supernatant and a cell pellet which was resuspended in RPMI + 10% FCS. The total number of nucleated cells in both organs for each mouse was counted with an electronic particle counter (Coulter Electronics, Hialeah, FL, USA). Blood was obtained by exsanguination of mice via cardiac puncture after administering the sedative, Atrovet (1 mg/ml), as 0.05 ml per 20. mg mouse, intraperitoneally, followed by the anesthetic, ketamine hydrochloride (100 mg/ml), administered as 0.02 ml, intramuscularly. Assay of NK cell functional activity
A radiolabelled chromium (51Cr) release assay, well standardized in our laboratory and others, was used with YAC-1 radiolabelled cells as targets (T) and spleen cells as effectors (E), in various E:T ratios from 3 1/8 to 100:1. Viability of all cells was at least 90% in all assays. The resulting cytotoxicity index (CI), after a 4 h co-incubation of E and T cells at each ratio was calculated as: CI =
cpm (experimental) - cpm (spontaneous release)
cpm (maximum release) - cpm (spontaneous release)
X
100,
where spontaneous, or non-specific, release is the radioactivity measured in cpm released in controls consisting of 50 JlI of 5x104 51Cr-Iabelled YAC-1 cells in medium only (200 JlI), and maximum release represents the same number of tumor cells lysed completely by incubation with 200 JlI detergent Zaponin (Coulter Electronics, Hialeah, FL, USA).
Strain differences in NK cell immunity . 27 NK cell identification and enumeration
The ASGM-1 surface molecule, a glycosphingolipid, is found on 100% of mature and maturing NK cells (48-51), all of which are morphologically distinct as being non-blast, small or medium lymphocytic cells (4, 5, 24-26, 44, 45, 52). In the present study, the ability to selectively identify, on hematologically counter-stained preparations, the ASGM-l surface marker on cells of typical and specific morphology, eliminated the possibility that any ASGM-1+cell not fitting the typical NK cell description (ASGM-l+, non-blast, lymphocytic), would be included in the NK cell enumeration process. To label the ASGM-l surface marker, spleen and bone marrow cells were incubated in suspensions (0 °C, 30 min) sequentially with a 1:40 concentration of rabbit anti-ASGM-l antibody, Lot # CAJ7837 (Wako Chemicals, Inc., Dallas, TX, USA), and biotinylated goat anti-rabbit IgG (1:100) (ABC Vectastain, Vector Laboratories, Burlingame, CA, USA). Anti-ASGM-l originated from bovine brain tissue repeatedly immunized with methylated BSA and complete Freund's adjuvant. The gamma globulin fraction of serum was obtained by 50% ammonium sulfate precipitation methods followed by dialyzing with PBS (pH 7.2), as described by Wako Chemicals, Inc. The cell suspensions were then cytocentrifuged, fixed in 100% methanol, exposed to 3.0% H 20 2 (10 min) to block endogenous peroxidase activity, treated with avidin-biotin-peroxidase complex (ABC Vectastain) and exposed to diaminobenzidine (DAB) for 13 min. ASGM-l+ cells were recognized by the presence of a granular DAB reaction produced at the cell periphery. In control experiments, cells were suspended with normal rabbit serum in place of rabbit antiASGM-l antibody. Cytospots were then counterstained with MacNeal's tetrachrome stain thereby permitting clear morphological identification of all ASGM-l+ cells. The proportions of ASGM- 1+ cells of non-blast, lymphocytic morphology, presumptive NK cells, were obtained from differential counts of 1,000 total lymphocytic cells (excluding blasts) per organ (spleen, bone marrow) per mouse and the results converted via the known total numbers of lymphocytic cells/organ to absolute numbers of NK cells per spleen or per femur. Isolation of blood lymphocytes
Freshly collected blood (1 ml), from mice of each strain, experimental and control, was placed in heparinized tubes with an equal volume of balanced salt solution. The diluted blood was then layered over 3 ml Ficoll-Paque (Pharmacia Biotech, Uppsala, Sweden) and centrifuged at 400xg for 40 min at 20°C. The distinct layer of lymphocytic cells was removed and suspended in RPMI 1640 + 10% FCS and centrifuged for 10 min at 4 °C to obtain a pellet which was then resuspended in 300 pi of cytospotting medium. These cells were deposited onto albumin-coated slides by means of a cytocentrifuge (Cytospin 2, Shandon, Inc., Montreal, QC, Canada). The slides were then immunoperoxidase-labelled and counterstained, as above, to permit the identification of NK cells. Isolation of NK cells by means of immunomagnetic beads
Spleen cells were incubated with the primary antibody (anti-ASGM-l) for 30 min on ice. The cells were washed twice subsequently by centrifugation, resuspended in medium (RPMI + 10% FCS) and incubated with the secondary antibody (goat anti-mouse IgG), covalently bound to immunomagnetic beads of 4.5 pm diameter (Dynabeads, Dynal, Inc., Oslo, Norway), for 30 min on ice. The tube containing the cells was applied to a magnetic particle concentrator (Dynal, Inc., Oslo, Norway) for 2 min followed by the removal of the supernatant. The purity of NK cells among the isolated cells ranged from 65 to 75% of all cells in the isolates. The cells attached to the beads were washed twice, using this method, resuspended in 1 ml medium + 15 pI Detachabead (Dynal, Inc., Oslo, Norway) a polyclonal antibody directed against the Fab fragments of mouse monoclonal antibodies, for 1 h at room temperature. The magnet was applied to the tube for 2 min and the isolated cells, detached from the beads, were collected in the supernatant. The isolated cells were then immunolabelled and counterstained as above.
28 . A. L. WHYTE and S. C. MILLER Statistics The unpaired two-tailed student t-test was used to determine the significance of the differences between means for the various experimental and control groups, within or between strains, where indicated. Values of p < 0.05 were considered significant.
Results Mice of the CS7BI/6 strain had significantly more nucleated cells in both their spleens and bone marrow than did the comparable organs from AI] mice (Table 1). Moreover, CS7Bl/6 mice also had significantly more (p < O.OS) lymphocytic cells in their spleens than did AI] mice, although the lymphoid cells in the bone marrow of CS7Bl/6 mice were not significantly elevated over those of AI] mice. The absolute number of natural killer (NK) cells, when assessed in both the spleen and bone marrow of the two strains, was found to differ in the spleen (wherein CS7Bl/6 mice had significantly more NK cells than AI] mice), but not in the bone marrow between the two strains (Fig. 1). Functional assays measuring the lytic ability of mature NK cells in both organs of the two strains, indicated that NK cells from the spleens of AI] mice had a significantly lower level of tumor-lytic ability at all NK cell-target cell ratios tested (p < O.OS-o.OOl) than did the splenocytes from CS7B1I6 mice (Fig. 2). Moreover, the NK cell frequency in blood of CS7B1I6 mice was significantly higher (p < O.OS) than that of AI] mice (Fig. 3). Table 2 illustrates the effect of in vivo administration of the powerful NK cell stimulant, ATRA, on the absolute numbers of NK cells. The results indicate no significant difference, in either the bone marrow or spleen, in the absolute numbers of NK cells between ATRA-treated and vehicle-administered mice within the same strain. By contrast, however, both the powerful NK stimulants, indomethacin and interleukin-2, significantly elevated the absolute numbers of
Table 1. Total cellularity and total numbers of lymphocytic cells in the spleen and bone marrow of A/] and C57B1I6 mice. Organ
Mouse strain
Total nucleated cells (xl 06) (mean ± s.e.)
Total lymphocytic cells (xl 06)" (mean ± s.e.)
Spleen
C57B1I6 AI]
110.3 ± 2.6 87.4 ± 2.8"-
86.9 ± 3.2 70.2 ± 5S:-
Bone marrow
C57B1I6 AI]
19.6 ± 0.5 15.0 ± OS'
5.2 ± 0.5 4.0 ± 0.8
Mean ± SE: 25-35 mice/strain. ':- p < 0.05 ':- Determined from differential counts of 1000 nucleated cells and converted via the known total cellularity to absolute numbers per spleen, or per femur.
Strain differences in NK cell immunity . 29 6
..
5
.! 'i
4
"'c
~
.."+::!
3
(/)
2
•
C57BII6
•
AlJ
u
0
~ ~
z
0
Spleen
Bone Marrow
Figure 1. The absolute numbers of NK(ASGM-l+) cells in the spleen and bone marrow (one femur) of C57B1I6 and A/] mice. The NK(ASGM-l+) cells in the spleen, but not the bone marrow of C57B1I6 mice are significantly more numerous (p < 0.05) than that of the spleen of A/] mice. Mean ± SE: 6-8 mice/group.
30..---------------------, --
C57BII6
- - - AlJ
0+---.------,,---...---........----.-----..---1 1 :3
1:6
1:12
1 :25
1 :50
1:100
Figure 2. Percent specific lysis of YAC-l lymphoma target cells by spleen cells from C57B1I6 and A/] mice. Mean ± SE: 5-7 mice/E:T ratio.
NK cells in both the spleen and the bone marrow within each strain, relative to vehicle-administered controls (Table 3). Figures 4 and 5 indicate the proportions of lymphocytic cells, in the spleens and bone marrow of both strains, whose immediate precursors had incorporated tritiated thymidine, a marker of DNA synthesis, at 1 h or 5 h earlier. At 1 h (Fig. 4) after a single injection of the radio-labelled DNA precursor, the proportions of lymphocytes bearing the isotope in the spleen and bone marrow, were similar between the two strains. Even at 5 h after isotope administration (Fig. 5), there
30 . A. L. WHYTE and S. C. MILLER Table 2. Effects of ATRA on the absolute numbers of NK(ASGM-1+) cells in the spleen and bone marrow of A/} and C57B1I6 mice. Organ
Mouse strain
Treatment
Absolute number of NK(ASGM-1+) cells (x1Q6)"
Spleen
C57B1I6
None b Corn oil vehiclec ATRAd
07 ± 0.44e 4.18 ± 0.22 3.74 ± 0.24
Bone marrow
C57B1I6
None Corn oil vehicle ATRA
0.45 ± 0.07 0.34 ± 0.04 0.29 ± 0.05
Spleen
A/}
None Corn oil vehicle ATRA
2.88 ± 0.32 2.40 ± 0.34 3.10 ± 0.26
Bone marrow
A/}
None Corn oil vehicle ATRA
0.36 ± 0.08 0.26 ± 0.03 0.29 ± 0.04
"Determined from differential counts of 1000 nucleated cells and converted via the known total nucleated cell count to absolute numbers per spleen, or per femur. bNo treatment. cCorn oil was given orally by gavage once/day for 5 days. dAll trans retinoic acid was administered (5 mg/kg) once/day for 5 days. eMean ± SE: 6-8 mice/group.
4
...... ~ ......
3
::IE CJ
2
...+•
rn c(
•
C57BV6
IIJJI AlJ
lie: Z
0-1-----
blood
Figure 3. NK(ASGM-t+) cells as a proportion of all lymphocytes in the blood of C57B1I6 and A/] mice, determined from differential analyses of 1000 total lymphocytes per hematologically stained cytospot. Mean ± SE: 6-8 mice/group. p < 0.05.
Strain differences in NK cell immunity .
31
Table 3. Effects of indomethacin or interleukin-2 on the total numbers of NK(ASGM-1+) cells in AI] and C57Bl/6 mice. Organ
Mouse strain
Treatment
Absolute numbers of NK(ASGM-1+) cells (x106)a
Spleen
C57B1I6
vehicle b indomethacin< interleukin-2d
5.4 ± O.4 e 7.2 ± 1.6 13.9 ± 3.1"
Bone marrow
C57B1I6
vehicle indomethacin interleukin-2
0.4 ± 0.04 1.0 ± 0.05':1.4 ± 0.2"-
Spleen
AI]
vehicle indomethacin interleukin-2
2.8 ± 0.1 3.9 ± 0.6" 6~4 ± 0.9"
Bone marrow
AI]
vehicle indomethacin interleukin-2
0.4 ± 0.03 0.9 ± 0.08'" 1.5 ± 0.4"
aDetermined from differential counts of 1,000 nucleated cells and converted via the known total nucleated cell count to absolute numbers per spleen or per femur. bThe vehicle for indomethacin is 1 ml absolute ethanol in the drinking water; the vehicle for interleukin-2 is 0.1 ml sterile phosphate buffered saline, containing 1% homologous serum, injected i.p. 3x/day for 4 days. <5 mg dissolved in 1 ml absolute ethanol, diluted in the drinking water to a final concentration of 10 jlg/ml. dRecombinant, human, given as bolus i.p. injections 3x/day for 4 days, each injection containing 12x106 units interleukin-2. eMean ± SE: 8 mice. "-p < 0.05-0.001; relative to vehicle-treated.
15 'C
.!! Gi
.c .!! CD
c
:c
E>.c
":"
12 •
C57BII6
fIrIII A1J
9 6
~
'#.
3
0
Spleen
Bone Marrow
Figure 4. Tritiated thymidine labelling index of lymphocytes in the spleen and bone marrow of C57Bl/6 and AI] mice. The in vivo time availability of the isotopically labelled, DNA precursor was 1 h - at which time the organs were sampled. Mean ± SE: 6-8 mice/group.
32 . A. L. WHYTE and S. C. MILLER Table 4. Effects of ATRA on the absolute numbers of 3H-thymidine-labelled lymphocytic cells in the spleen and bone marrow of AI] and C57B1I6 mice. Organ
Mouse strain
Treatment
Absolute numbers of H-thymidine labelled cells (x10 6).,b
Spleen
C57Bl/6 C57Bl/6
Corn oil vehiclec ATRAd Corn oil vehicle ATRA
6.74 ± 1.46< 5.57 ± 0.30 3.82 ± 0.50 3.90 ± 0.50
C57B1I6 C57Bl/6
Corn oil vehicle ATRA Corn oil vehicle ATRA
1.40 ± 0.14 1.45 ± 0.14 1.06 ± 0.19 0.89 ± 0.07
A/] AI] Bone Marrow
AI] AI]
'Determined from differential counts of 1000 nucleated cells and converted via the Known total nucleated cell count to absolute numbers per spleen, or per femur. bH-thymidine was given as a single pulse injection, i.p. (1 pCi/mg body wt), at the conclusion of 5 days of ATRA administration in the diet. cCorn oil was given orally by gavage once/day for 5 days. dAll trans retinoic acid was administered (5 mg/kg) once/day for 5 days.
was no significant difference in the proportions of isotope-bearing lymphocytes when comparing the bone marrow organs between the two strains, or, when comparing the spleens between the two strains. Within each strain, the proportion of labelled lymphocytes in the bone marrow at 5 h was approximately 3-fold that observed at 1 h.
15 'tI
.!! 'ii .c
12
"c
9
••
.!!
'6
C57BII6
AlJ
E
>-
i
~
6
..,
.,.
3 0
Spleen
Bone Marrow
Figure 5. Tritiated thymidine labelling index of lymphocytes in the spleen and bone marrow of C57B1I6 and A/] mice. The in vivo time availability of the isotopically labelled, DNA precursor was 5 h - at which time the organs were sampled. Mean ± SE: 6-8 mice/group.
Strain differences in NK cell immunity . 33
2r-------------------, •
II
CS7B1/6 AlJ
-
'",.. o
MEL·14
MAC·1
Figure 6. Absolute numbers of lymphocytes (enriched for NK cells) in the spleen which bear the surface molecule recognized by mAb MEL-14 or Mac-I. Mean ± SE: 5 mice/group.
At the 5 h interval, however (Fig. 5), there was a 52% reduction (p < 0.001) in the proportion of labelled lymphocytes in AI] spleen vs that of AI] bone marrow, whereas, by contrast, the proportions of labelled lymphocytes in the spleen and the bone marrow of C57Bl/6 mice were almost identical. Table 4 indicates that no significant difference, either strain-dependent or organ-dependent, could be found in the absolute numbers of labelled lymphocytes when mice were injected with ATRA or its vehicle. Finally, Figure 6 illustrates the absolute numbers of splenic lymphocytes, highly enriched for NK cells after Dynabead separation, which bore the homing receptor, gp90MEL-t4 or the integrin, Mac-t. The absolute numbers of these Dynabead-enriched lymphocytes, the majority of which are NK cells, which bore either of these surface molecules, differed either only marginally (gp90MEL-14), or insignificantly (MAC-t) between the two strains.
Discussion The low cellularity of the bone marrow, the primary production site of all the hemopoietic cells, of mice of the AI] strain relative to that of C57Bl/6 mice appears to reflect the previously demonstrated low levels of GM-CFC's in AI] bone marrow compared with C57Bl/6 mice (53). The consequence of this phenomenon would be, as the present data demonstrate, significantly lower numbers in AI] mice, of hemopoietic cells in the spleen, a major recipient organ of mature cells originating in the bone marrow. We have recently shown that, indeed, AI] mice have significantly lower absolute numbers of lymphoid, granulocytic and monocytic - but not erythroid - cells in their spleen andlor bone marrow than do C57Bl/6 mice (54).
34 . A. L. WHYTE and S. C. MILLER
It appears that not all of the NK(ASGM-1 +) cells produced in the bone marrow of the AI] mouse (whose numbers are similar to those of C57Bl!6] mouse bone marrow) are leaving the bone marrow, as reflected by the significantly lower proportions of these cells in AI] blood vs C57Bl!6 blood. The significantly lower NK(ASGM-1+) cell content of AI] spleen, would correspondingly reflect the low blood levels, the spleen being the primary destiny for virtually all bone marrow-derived, newly generated NK cells, and the blood, being the only route out of the bone marrow. The proportions of 3H-thymidine labelled lymphocytes were similar in the corresponding organs of the two strains at 1 h after giving a single in vivo injection of the radioisotope, indicating that proliferation of the lymphoid precursors, as a whole, was comparable in the two strains, in spite of significantly lower levels of NK cells (and other lymphoid cells) in the spleen. At 5 h after the single injection of labelled nucleotide (tritiated thymidine), C57Bl!6 mouse bone marrow and spleen (a major recipient of bone marrow-produced Band NK cells), contained equivalent proportions of isotope labelled lymphoid cells. The observation that there was no decrease, and in fact, some increase, in the numbers of lymphoid cells in the bone marrow of C57B1/6] mice at 5 h (relative to the 1 h value for that organ), simply indicates the mitotic activity of those cells which incorporated the isotope at 1 h. The 2 daughter cells, at 5 h now each with it'~ own complement of isotopic grains derived from the par'ent cell, have added to the recordable (countable) numbers of cells «labelled» at 5 h. Furthermore, free isotope available in vivo, may be picked up and utilized by any such cells going into DNA-synthesis (and subsequently mitosis). These lymphoid cells would also contribute to the numbers of labelled cells at 5 h. However, there appears in AI] mouse bone marrow to be, as well, a considerable retention of newly produced (isotope labelled) lymphoid cells, such that the splenic proportions of these cells are less than half those of the bone marrow of that strain at 5 h post thymidine injection. Significantly lower numbers of NK cells in AI] blood, the only exit route from the bone marrow, further support retention of at least some of these cells in that organ. These results collectively suggest a level of production of NK cells in AI] mice equivalent at least to that of C57Bl6 mice, but, apparently, one or more phenomena are taking place: (i) considerable postproduction intramyeloid NK cell deathlabortion is occurring, (ii) many NK cells are themselves intrinsically incapable of emigrating from the bone marrow, or, (iii) NK cells are being inhibited by the endogenous myeloid-infra-structure, from leaving that organ. A low level of bone marrow-to-spleen trafficking of newly formed lymphocytes, as a whole, in AI] mice, may account not only for the significantly lower NK cell frequency in AI] blood (vs C57Bl!J) (and consequently, the lower absolute numbers of NK(ASGM-1+) cells observed in the AI] spleen vs C57Bl!7 spleen), but also for the lower total numbers of lymphocytes in AI] spleen. By contrast, the bone marrow of both strains, contains similar numbers of lymphoid cells, a fact, which together with the kinetic studies involving tritiated thymidine uptake, does indeed reflect similar production levels of lymphoid cells by the bone marrow of both strains. Band NK lymphoid cells are bone marrow-generated, virtually exclusively.
Strain differences in NK cell immunity .
35
Although ATRA has been shown to increase «NK activity», i.e., lysis of tumor cells (27), it was not known whether this ATRA-stimulated increase in lytic function may result from an increase in absolute number of NK cells in the presence of ATRA. The present results indicate, however, that even after 5 days of exposure to ATRA (spanning, thus, at least 4 NK cell generation cycles (13», no increment of NK cell numbers could be found, in either AI] or C57Bl6 mouse organs. This finding strengthens previous observations which have demonstrated distinct ATRA-enhanced subcellular events such as increases in perforin levels and mRNA levels in killer cells (29). By contrast, two other potent NK-stimulating agents, indomethacin and interleukin-2, have resulted, in both organs (spleen and bone marrow), of both strains of mice (AI] and C57B1I6), in significantly elevated numbers of NK cells, indicating the influence of these agents at strain (genetically) independent stages of new cell genesis. The influence of these agents is consistent with our previous observations in other strains of mice (40-43). The homing receptor gp90MEL-14, and the integrin, MAC-1, are essential to transendothelial migration of lymphocytes and other blood-borne cells. DYNABEAD separated splenocytes, enriched for NK cells, demonstrated no difference between the 2 strains in their capacity to bear the homing receptor andlor the integrin. These observations suggest equivalent competence of NK(ASGM1+) cells in both strains with respect to localizing at, and binding to, endothelial cell walls such as would be encountered prior to extravasation from splenic vascular channels into the splenic parenchyma, the site of NK cell function. In conclusion, it appears that the long-established «low NK cell-mediated activity» of mice of the AI] strain (vs the «high NK cell-mediated activity» of the C57B1I6 strain) may have a single cause, i.e., lack of substantial exit (because of intra-myeloid abortion andlor inhibition of emigration) of newly produced, NK cell progeny from their bone marrow birth site. References 1. TIMONEN, T., A. RANKL, E. SAKSELA, and P. HAYRY. 1979. Human cell-mediated cytotoxicity against fetal fibroblasts. III. Morphological and functional characterization of the effector cells. Cell. Immunol. 48: 121. 2. KUMAGAI, K., K. ITOH, R. SUZUKI, S. HUNUMA, and F. SAITOH. 1982. Studies of murine large granular lymphocytes. I. Identification as effector cells in NK and K cytotoxicities. J. Immunol. 129: 388. 3. POLLACK, S. B., and C. ROSSE. 1987. The primary role of murine bone marrow in the production of NK cells: a cytokine study. J. Immunol. 139: 2149. 4. KIESSLING, R., E. KLEIN, H. PROSS, and H. WIGZELL. 1975. Natural killer cells in the mouse. II. Cytotoxic cells with specificity for mouse moloney leukemic cells: characteristics of the killer cell. Eur. J. Immunol. 5: 117. 5. BIRON, C. A., and R. M. WELSH. 1982. Activation and role of natural killer cells in virus infections. Med. Microbiol. Immunol. 170(3): 155. 6. TALMADGE, J. E., K. M. MEYERS, D. J. PRIEUR, and J. R. STARKEY. 1980. Role of NK cells in .tumor growth and metastasis in beige mice. Nature 284: 622. 7. HANNA, N., and R. C. BURTON. 1981. Definitive evidence that natural killer cells inhibit experimental tumor metastasis in vivo. J. Immunol. 127: 1754.
36 . A. 1. WHYTE and S. C. MILLER 8. JANEWAY, C. A. 1989. A primitive immune system. Nature 341: 108. 9. HALLER, O. R. KIESSLING, A. ORN, and H. WIGZELL. 1977. Generation of NK cells: An autonomous function of bone marrow. J. Exp. Med. 145: 1411. 10. HACKETT, J., M. BENNETT, G. C. Koo, and V. KUMAR. 1986. Origin and differentiation of natural killer cells. ImmunoI. Res. 5: 16. 11. KALLAND, T. 1986. Generation of natural killer cells from bone marrow precursors in vitro. ImmunoI. 57: 493. 12. POLLACK, S. B., M. Y. LEE, E. COHEN, J. 1. LOTTSFELDT, and C. ROSSE. 1989. Modulation of murine natural killer cells by a granulocytosis-inducing tumor. Cancer Res. 49: 174. 13. MILLER, S. C. 1982. Production and renewal of murine NK cells in the spleen and bone marrow. J. ImmunoI. 129: 2282. 14. GREENBERG, A. H., and M. GREEN. 1979. Non-adaptive rejection of small tumor inocula as a model of immune surveillance. Nature 264: 356. 15. ZOLLER, M., D. BELLGRAW, I. AxBERG, and H. WIGZELL. 1982. Natural killer cells do not belong to the recirculating lymphocyte population. Scand. J. ImmunoI. 15: 159. 16. Ross, G. D., and V. VETVICKA. 1993. CR3 (CDllb/CD18) a phagocyte and NK cell membrane receptor with multiple ligand specificities and functions. Clin. Exp. ImmunoI. 192: 181. 17. SOMERSALO, K., J. TARKKANEN, M. PATARROYA, and F. SAKSELA. 1992. Involvement of ~2 integrins in the migration of human natural killer cells. J. ImmunoI. 149(2): 590. 18. DIAMOND, M. S., D. E. STAUNTON, A. R. FOUGEEROLLES, S. A. STACKER, J. GARCIAAGUILAR, M. C. GIBBS, and T. A. SPRINGER. 1990. ICAM-1 (CD54): A counter-receptor for Mac-1 (CDllb/CD18). J. Cell BioI. 111: 3129. 19. SMITH, C. W., S. D. MARLIN, R. ROTHLEIN, C. TOMAN, and D. C. ANDERSON. 1989. Cooperative interactions of LFA-1 and MAC-1 with intercellular adhesion molecule-1 in facilitating adherence and transendothelial migration. J. Clin. Invest. 82: 2008. 20. DIAMOND, M. S., J. GARCIA-AGUILAR, J. K. BICKFORD, A. 1. CORBI, and T. A. SPRINGER. 1993. The I domain is a major recognition site on the leukocyte integrin Mac-1 (CDIIb/ CD18) for four distinct adhesion ligands. J. Cell. BioI. 120: 1031. 21. BUTCHER, E. C. 1991. Leukocyte-endothelial cell recognition: three (or more) steps to specificity and diversity. Cell 67: 1033. 22. MILLER, S. C. 1994. The development of natural killer (NK cells) from Thy-1(10)Lin-Sca-1+ stem cells: Acquisition by NK cells in vivo of the homing receptor MEL-14 and the integrin MAC-I. ImmunobioI. 190(4-5): 385. 23. KISHIMOTO, T. K., M. A. JUTILA, and E. C. BUTCHER. 1990. Identification of a human peripheral lymph node homing receptor: A rapidly down-regulated adhesion molecule. Proc. NatI. Acad. Sci. USA 87: 2244. 24. CHRISTOPHER, F. 1., I. DUSSAULT, and S. C. MILLER. 1991. Population dynamics of natural killer cells in the spleen and bone marrow of normal and leukemic mide during in vivo exposure to interleukin-2. ImmunobioI. 184: 37. 25. MILLER, S. c., F. 1. CHRISTOPER, and I. DUSSAULT. 1992. Population dynamics of natural killer cells and T lymphocytes in murine spleen and bone marrow during the development of erythroleukemia: The effect of indomethacin. Nat. Immun. 11: 78. 26. DUSSAULT, I., and S. C. MILLER. 1993. Stimulation of natural killer cell numbers but not function in leukemic infant mice: A system primed in infancy allows survival in adulthood. Nat. Immun. 12: 66. 27. GOETTSCH, W, Y. HATORI, and R. SHARMA. 1992. Adjuvant activity of all-trans-retinoic acid in C57Bl/6 mice. Immunopharma. 14(2): 143. 28. GOLDFARB, R. H., and R. B. HERBERMAN. 1981. Natural killer cell reactivity: regulatory interactions among phorbol ester, interferon, cholera toxin and retinoic acid. J. ImmunoI. 126(6): 2129. 29. LIN, T. H., and T. M. CHU. 1993. Enhancement of perforin by retinoic acid is mediated by protein kinase C. Biochem. And Biophys. Res. Comm. 191(3): 937, 30. FEGAN, c., BAILEy-WOOD, S. COLEMAN, S. A. PHILLIPS, 1. NEALE, T. Hoy, and J. A.
Strain differences in NK cell immunity . 37 WITI'AKER. 1995. All trans retinoic acid enhances human LAK activity. Eur. J. Haematol. 54: 95. 31. HERBERMAN, R. B., M. E. NUNN, and D. H. LARVIN. 1975. Natural cytotoxic reactivity of mouse lymphoid cells against syngeneic and allogeneic tumors. I. Distribution of reactivity and specificity. Int. J. Cancer 16.216. 32. HALLER, O. A., M. GIDLUNG, J. T. KURNICH, and H. WIGZELL. 1978. In vivo generation of mouse natural killer cells: role of spleen and thymus. Scand. J. Immunol. 8: 207. 33. KAMINISKY, S. G., I. NAKAMURA, and G. CUDKOWICZ. 1983. Selective defect of natural killer and killer cell activity against lymphomas in SJL mice: Low responsiveness to interferon inducers. J. Immunol. 130: 1980. 34. ITOH, K., R. SUZUKI, Y. UMEZU, et al. 1982. Studies of murine large granular lymphocytes. II. Tissue, strain and age distributions of LGL and LAL. J. Immunol. 129: 395. 35. CHEERS, c., and I. F. C. McKENZIE. 1978. Resistance and susceptibility of mice to bacterial infection: genetics of listeriosis. Infect. Immun. 19: 755. 36. PENNINGTON, J. E., and R. M. WILLIAMS. 1979. Influence of genetic factors on natural resistance of mice to Pseudomonas aeruginosa. J. Infect. Dis. 139: 396. 37. BROWN, I. N., A. A. GLYNN, and J. PLANT. 1982. Inbred mouse strain resistance to Myobacterium lepraemurium follows the Ity/Lsh pattern. Immunology 47: 149. 38. NESBITI', M. N., and E. SKAMENE. 1984. Recombinant inbred mouse strains derived from A/J and C57B1I6J: A tool for the study of genetic mechanisms in host resistance to infection and malignancy. J. Leuk. BioI. 36: 357. 39. YOSHIDA, A., Y. KOIDE, M. UCHIJIMA, and T. O. YOSHIDA. 1995. Dissection of strain difference in acquired protective immunity against Mycobacterium bovis Calmette-Guerin Bacillus (BCG). Macrophages regulate the susceptibility through cytokine network and the induction of nitric oxide synthetase. J. Immunol. 155: 2057. 40. ROSENSTEIN, M., S. E. ETTINGHAUSEN, and S. A. ROSENBERY. 1986. Extravasation of intravascular fluid mediated by the systemic administration of recombinant interleukin-2. J. Immunol. 137: 1735. 41. KOHLER, P. c.,J. A. HANK, K. H. MOORE, B. STORER, R. BECHHOFER, R. HONG, and P. M. SONDEL. 1989. Phase I clinical trial of recombinant interleukin-2: A comparison of bolus and continuous intravenous infusion. Cancer Invest. 7: 213. 42. WEST, W. H., K. W TAUER, J. R. YANNELLI, G. D. MARSHALL, D. W. ORR, G. B. THURMAN, and R. K. OLDHAM. 1987. Constant infusion recombinant interleukin-2 in adoptive immunotherapy of advanced cancer. N. Engl. J. Med. 316: 898. 43. GATELY, M. K., T. D. ANDERSON, and T. J. HAYES. 1988. Role of Asialo-GMI-positive lymphoid cells in mediating the toxic effects of recombinant IL-2 in mice. J. Immunol. 141: 189. 44. MILLER, S. c., and A. C. SHATZ. 1991. Relationship between large and small tumor-binding cells in the spleen and bone marrow. Nat. Immun. & Cell Growth Reg. 10: 320. 45. DUSSAULT, I., and S. C. MILLER. 1994. Decline in natural killer cell mediated immunosurveillance in aging mice - a consequence of reduced cell production and tumor-binding capacity. Mech. Ageing & Dev. 75: 115. 46. WAHL, L. M., and L. LAMPEL. 1987. Regulation of human peripheral blood monocyte collagenase by prostaglandins and anti-inflammatory drugs. Cell. Immunol. 105: 411. 47. GOLDMAN, R. 1984. Effect of retinoic acid on the proliferation and phagocytic capability of muring macrophage-like cell lines. J. Cell. Physiol. 120: 91. 48. KASAl, M., M. IWAMORI, #. NAGAI, K. OKUMURA, and I. TADA. 1980. A glycolipid on the surface of mouse natural killer cells. Eur. J. Immunol. 10: 175. 49. YOUNG, W W, S. HAKOMORI, J. M. DURDIK, and C. S. HENNEY. 1980. Identification of ganglio-N-temiosylceramide as a new surface marker for murine natural killer (NK) cells. J. Immunol. 124: 199. 50. STOUT, R. D., G. A. SCHWARTING, and J. SUTTLES. 1987. Evidence that expression of AsialoGm1 may be associated with cell activation. J. Immunol. 139: 2123. 51. BECK, B. N., S. GILLIS, and C. S. HENNEY. 1982. Display of the neutral glycolipid ganglio-
38 . A. L. WHYTE and S. C. MILLER N-tetraosylceramine (Asialo Gm1) on cells of the natural killer and T lineages. Transplantation 33: 118. 52. RODER, J. c., and R. KIESSLING. 1978. Target-effector interaction in the natural killer cell system. I. Covariance and genetic control of cytolytic and target-cell binding subpopulations in the mouse. Scand.]. Immunol. 8: 135. 53. POZZULO, G. N., E. SKAMENE, and F. GERVAIS. 1993. Bone marrow cell response following induction of acute inflammation in different strains of mice. Inflammation 17(6): 677. 54. MILLER, S. c., and S. L. KEARNEY. 1998. Effect of in vivo administration of trans retinoic acid on the hemopoietic cell populations of spleen and bone marrow: Profound strain differences between AI] and C57B1I6] mice. Lab. Animal Sci. 48 (1): 14. SANDRA C. MILLER, Ph.D., Department of Anatomy and Cell Biology, McGill University, 3640 University Ave, Room 2/28, Montreal, Quebec, Canada H3A 2B2
Department of Anatomy and Cell Biology, McGill University, Montreal, Quebec, Canada
Strain Differences in Natural Killer Cell-mediated Immunity among Mice: A Possible Mechanism for the Low Natural Killer Cell Activity of AIJ Mice ANDREA L. WHYTE and SANDRA
C. MILLER
Received July 7, 1997 . Accepted in revised form November 20, 1997
Abstract Natural killer (NK) cells are well established as fundamental elements in the early eradication of aberrant cells potentially leading to neoplasia. Moreover, it has also long been known that inbred strains of laboratory mice, as well as human individuals, demonstrate a wide range of NK cellmediated immune response even to the same tumor. In the present study, various parameters which could lead ultimately to high, or low, NK cell-mediated functional activity have been assessed. Mice of the A/] strain demonstrate very low NK cell tumor-lytic activity and correspondingly high incidence of lymphoma. By contrast, C57B1I6 mice demonstrate relatively high NK cell activity and virtually never develop lymphomas. The results of this study have revealed that the absolute numbers of splenic NK cells were significantly lower in A/] vs C57BI6/mice. Furthermore, the blood of A/] mice contained significantly fewer (30%) NK cells than did that of C57B1I6 mice. However, no significant difference between the 2 strains was found in the numbers of lymphocytic cells from NK cell-enriched fractions from the spleens, which possessed either the homing receptor MEL-H, or the integrin Mac-I, both essential surface molecules for transendothelial migration of lymphocytic cells from the circulation into organ parenchyma. Moreover, NK cells from both strains responded similarly to the NK cell stimulants, ATRA, indomethacin and interleukin-2. Finally, there was no significant difference between the 2 strains, in the numbers of lymphocytic cells in the bone marrow (including NK cells), which were radiolabelled with the DNA synthetic precursor, 3H-thymidine, indicative, thus, of equivalent levels of lymphocyte production by the bone marrow in the 2 strains. The observations collectively suggest that the low peripheral (spleen, blood) levels of NK cell-mediated functibnal activity found in the A/] strain of mouse at least, reflects either post-production, large-scale NK cell abortion/death, or a bone marrow-based microenvironmental deficiency which inhibits NK cells' exit from the bone marrow birth site.
Abbreviations: ATRA = All trans retinoic acid, rIl-2 - recombinant interleukin-2; ASGM-l'= Asialo-GM-1; DAB = diaminobenzidine, CI = cytotoxicity index; NK = natural killer; FCS = fetal calf serum. C1998
by Gustav Fischer Verlag
24 . A. L. WHYTE and S. C. MILLER
Introduction Natural killer (NK) cells have been characterized morphologically as granular, non-blast lymphocytic cells, small to medium in size, in all species studied (1-3). NK cells have the ability to lyse a wide variety of abnormal cells, i.e., tumor and virus-infected cells, and in vivo, there is little doubt that NK cells provide the primary defence or immunosurveillance against the spontaneous development of tumors (4-8). The origin of NK cells is the bone marrow (9-12). Newly-formed, mature NK cells are not themselves in cell cycle and leave the bone marrow birth site, randomly after their production (13). NK cells have an intramyeloid T t/2 of 8 h, followed by a circulation (blood transit) time of approximately 10 h while they are en route to the spleen wherein the T t/2 of NK cells is 24.5 h (13). Longevity, memory and recirculation are not characteristics of NK cells (13-15). The integrin, Mac-1 (CDllb/CD18), found on NK cells has been implicated, at least in vitro, for transendothelial migration, where it complexes with the endothelial ligand, ICAM-1 (16-20). The homing receptor gp90MEL-14 (Lselectin) permits loose, initial binding of lymphocytes with endothelium (21), and the receptor has recently been found on NK cells in vivo (22), suggesting a physiological role on NK cells, for this surface molecule long known to be essential for in vivo transendothelial migration of T and B lymphocytes. The homing receptor is soon shed while other, more firmly adhering, adhesion molecules such as Mac-1 are up-regulated to permit a tight lymphocyte-endothelial cell union immediately prior to transendothelial migration (23). There are 3 known potent stimulators of NK cell functional activity: the cytokine, interleukin-2, the drug, indomethacin and the metabolite of vitamin A, all-trans retinoic acid (ATRA). We have shown previously in mice which normally express relatively high levels of NK cell-mediated functional activity, that indomethacin and interleukin-2 augment the absolute numberslproduction of NK cells, leading, thus, to elevated functional levels of this cell in any organ tested (24-26). ATRA has been shown to enhance NK cell functional activity in mice and humans by up-regulating TNF-
Strain differences in NK cell immunity . 25
Collectively, the results have shown that, indeed, mature NK cells in AI] mice, are significantly lower in absolute numbers than in C57Bl/6 mice. Moreover, the NK cell numbers in both C57Bl/6 and AI] mice could be enhanced in vivo by interleukin-2, indomethacin but not by ATRA. However, lymphocyte production (NK and B cells) by the bone marrow (assessed by the frequency of lymphoid cells bearing a radiolabelled nucleotide precursor of DNA-s phase) was similar in both strains contrasting, thus, with the significantly lower levels of NK cells in the blood and spleen of AI] mice, and also with the significantly lower levels of all lymphoid cells found in the spleens of AI] relative to C57B1I6 mice. In AI] mice, thus, there appears to be a significantly lower emigration of newly generated NK cells (relative to C57B1I6 mice) from the bone marrow.
Materials and Methods Mice
A/J and C57B1I6 male mice (6-9 wk) were obtained from the Jackson Laboratories, Bar Harbor, ME, USA. Mice were housed in microisolator cages, maintained on standard complete mouse chow diet and remained undisturbed for 1 week prior to commencement of any experimentation. Tumor cell line
YAC-l, a lymphoma originating from A/Sn mouse strain, maintained in suspension culture consisting of RPMI 1640 medium with 10% FCS and 1% antibiotics and fungizone (GIBCO, Inc., Canada), served as the target of NK cells. Drug administration
Indomethacin (Sigma Chemical Co., St. Louis, MO, USA) was administered to mice through the drinking water. Five mg of the drug were dissolved in 1 ml absolute ethanol and diluted with water to give a concentration of 10 }lg/ml. Fresh solution was provided in drinking water every second day. No difference in the amount of water consumed between groups of mice given indomethacin, the indomethacin vehicle (ethanol in water), or untreated drinking water was observed as determined by measurements taken at 2 day intervals over the course of 1 wk. Cytokine administration
Human recombinant interleukin-2 (rl1-2), a generous gift from Hoffmann-LaRoche, Nutley, NJ, USA, was given as bolus intraperitoneal injections 3 times a day at 8 h intervals for 3 days. Recombinant 11-2, provided in the lyophilized state had a specific activity of 14xl06 Units/mg protein. A dose of 12xl03 Units in 0.1 rnl sterile phosphate-buffered saline, containing 1% homologous serum was administered at each injection. Toxicity levels, optimal dose-response curves, and injection frequency of rl1-2 in mice have been all well established by ourselves and others (24, 26, 40-43). Isotope administration
Tritiated thymidine, a radiolabelled precursor of DNA, readily incorporated by cells undergoing DNA synthesis, was given to mice intraperitoneally as a single dose of 1 }lCi/mg body weight, specific activity 20 Ci/rnM (ICN Biomedicals, Inc. CA, USA) in 0.5 rnl phosphate buffered
26 . A. L. WHYTE and S. C. MILLER saline. One h or 5 h later, mice were killed and cytospot samples of the spleen and bone marrow of 6-8 mice were processed for radioautography by the standard methods used in our laboratory (13, 44, 45). Administration of all-trans retinoic acid (ATRA)
ATRA (Sigma Chemical Co., St. Louis, MO, USA) was dissolved in commercially available corn oil (0.125 mg/0.1 ml), as the vehicle for oral administration immediately prior to administration, employing standard techniques of ATRA gavaging (46, 47). Five mg/kg body weight of ATRA was given every 24 h for 5 days. This time course was selected as more than adequate to ensure that any effect the agent might have on NK cell numbers/production would be readily evidenced given the short life span and rapid renewal of NK cells (13-15). In all cases, control mice were fed the vehicle only. Identification of homing receptor and integrin molecules on NK cells
Rat anti-mouse CDllb/CD18 (Mac-1) and rat anti-mouse L-selectin (MEL-14), purchased from Cedarlane Laboratories, ON, Canada), were diluted to optimal concentrations (based on dpse-response curves carried out in pilot experiments) of 1:5 and 1:30, respectively, prior to use. These primary antibodies were linked subsequently to biotinylated goat anti-rat IgG (Cedarlane Laboratories, ON, Canada) at concentrations of 1:100. Cells so stained were processed by the Vectastain method (Vector Laboratories, CA, USA), for identification of lymphocytes bearing either the homing receptor or the integrin. Tissue sampling
Spleens from experimental and control mice were removed and pressed through stainless steel screens into Petri dishes containing 5-10 ml ice-cold RPMI containing 10% FCS. From the same mice, the femurs were removed and flushed repeatedly with fresh RPMI + 10% FCS. Both spleen and bone marrow cell preparations were layered over 1 ml NCS for 5 min to allow sedimentation of large debris through the NCS. The cell suspensions were then removed, centrifuged for 5 min (500xg, 4 0c) to yield a supernatant and a cell pellet which was resuspended in RPMI + 10% FCS. The total number of nucleated cells in both organs for each mouse was counted with an electronic particle counter (Coulter Electronics, Hialeah, FL, USA). Blood was obtained by exsanguination of mice via cardiac puncture after administering the sedative, Atrovet (1 mg/ml), as 0.05 ml per 20. mg mouse, intraperitoneally, followed by the anesthetic, ketamine hydrochloride (100 mg/ml), administered as 0.02 ml, intramuscularly. Assay of NK cell functional activity
A radiolabelled chromium (51Cr) release assay, well standardized in our laboratory and others, was used with YAC-1 radiolabelled cells as targets (T) and spleen cells as effectors (E), in various E:T ratios from 3 1/8 to 100:1. Viability of all cells was at least 90% in all assays. The resulting cytotoxicity index (CI), after a 4 h co-incubation of E and T cells at each ratio was calculated as: CI =
cpm (experimental) - cpm (spontaneous release)
cpm (maximum release) - cpm (spontaneous release)
X
100,
where spontaneous, or non-specific, release is the radioactivity measured in cpm released in controls consisting of 50 JlI of 5x104 51Cr-Iabelled YAC-1 cells in medium only (200 JlI), and maximum release represents the same number of tumor cells lysed completely by incubation with 200 JlI detergent Zaponin (Coulter Electronics, Hialeah, FL, USA).
Strain differences in NK cell immunity . 27 NK cell identification and enumeration
The ASGM-1 surface molecule, a glycosphingolipid, is found on 100% of mature and maturing NK cells (48-51), all of which are morphologically distinct as being non-blast, small or medium lymphocytic cells (4, 5, 24-26, 44, 45, 52). In the present study, the ability to selectively identify, on hematologically counter-stained preparations, the ASGM-l surface marker on cells of typical and specific morphology, eliminated the possibility that any ASGM-1+cell not fitting the typical NK cell description (ASGM-l+, non-blast, lymphocytic), would be included in the NK cell enumeration process. To label the ASGM-l surface marker, spleen and bone marrow cells were incubated in suspensions (0 °C, 30 min) sequentially with a 1:40 concentration of rabbit anti-ASGM-l antibody, Lot # CAJ7837 (Wako Chemicals, Inc., Dallas, TX, USA), and biotinylated goat anti-rabbit IgG (1:100) (ABC Vectastain, Vector Laboratories, Burlingame, CA, USA). Anti-ASGM-l originated from bovine brain tissue repeatedly immunized with methylated BSA and complete Freund's adjuvant. The gamma globulin fraction of serum was obtained by 50% ammonium sulfate precipitation methods followed by dialyzing with PBS (pH 7.2), as described by Wako Chemicals, Inc. The cell suspensions were then cytocentrifuged, fixed in 100% methanol, exposed to 3.0% H 20 2 (10 min) to block endogenous peroxidase activity, treated with avidin-biotin-peroxidase complex (ABC Vectastain) and exposed to diaminobenzidine (DAB) for 13 min. ASGM-l+ cells were recognized by the presence of a granular DAB reaction produced at the cell periphery. In control experiments, cells were suspended with normal rabbit serum in place of rabbit antiASGM-l antibody. Cytospots were then counterstained with MacNeal's tetrachrome stain thereby permitting clear morphological identification of all ASGM-l+ cells. The proportions of ASGM- 1+ cells of non-blast, lymphocytic morphology, presumptive NK cells, were obtained from differential counts of 1,000 total lymphocytic cells (excluding blasts) per organ (spleen, bone marrow) per mouse and the results converted via the known total numbers of lymphocytic cells/organ to absolute numbers of NK cells per spleen or per femur. Isolation of blood lymphocytes
Freshly collected blood (1 ml), from mice of each strain, experimental and control, was placed in heparinized tubes with an equal volume of balanced salt solution. The diluted blood was then layered over 3 ml Ficoll-Paque (Pharmacia Biotech, Uppsala, Sweden) and centrifuged at 400xg for 40 min at 20°C. The distinct layer of lymphocytic cells was removed and suspended in RPMI 1640 + 10% FCS and centrifuged for 10 min at 4 °C to obtain a pellet which was then resuspended in 300 pi of cytospotting medium. These cells were deposited onto albumin-coated slides by means of a cytocentrifuge (Cytospin 2, Shandon, Inc., Montreal, QC, Canada). The slides were then immunoperoxidase-labelled and counterstained, as above, to permit the identification of NK cells. Isolation of NK cells by means of immunomagnetic beads
Spleen cells were incubated with the primary antibody (anti-ASGM-l) for 30 min on ice. The cells were washed twice subsequently by centrifugation, resuspended in medium (RPMI + 10% FCS) and incubated with the secondary antibody (goat anti-mouse IgG), covalently bound to immunomagnetic beads of 4.5 pm diameter (Dynabeads, Dynal, Inc., Oslo, Norway), for 30 min on ice. The tube containing the cells was applied to a magnetic particle concentrator (Dynal, Inc., Oslo, Norway) for 2 min followed by the removal of the supernatant. The purity of NK cells among the isolated cells ranged from 65 to 75% of all cells in the isolates. The cells attached to the beads were washed twice, using this method, resuspended in 1 ml medium + 15 pI Detachabead (Dynal, Inc., Oslo, Norway) a polyclonal antibody directed against the Fab fragments of mouse monoclonal antibodies, for 1 h at room temperature. The magnet was applied to the tube for 2 min and the isolated cells, detached from the beads, were collected in the supernatant. The isolated cells were then immunolabelled and counterstained as above.
28 . A. L. WHYTE and S. C. MILLER Statistics The unpaired two-tailed student t-test was used to determine the significance of the differences between means for the various experimental and control groups, within or between strains, where indicated. Values of p < 0.05 were considered significant.
Results Mice of the CS7BI/6 strain had significantly more nucleated cells in both their spleens and bone marrow than did the comparable organs from AI] mice (Table 1). Moreover, CS7Bl/6 mice also had significantly more (p < O.OS) lymphocytic cells in their spleens than did AI] mice, although the lymphoid cells in the bone marrow of CS7Bl/6 mice were not significantly elevated over those of AI] mice. The absolute number of natural killer (NK) cells, when assessed in both the spleen and bone marrow of the two strains, was found to differ in the spleen (wherein CS7Bl/6 mice had significantly more NK cells than AI] mice), but not in the bone marrow between the two strains (Fig. 1). Functional assays measuring the lytic ability of mature NK cells in both organs of the two strains, indicated that NK cells from the spleens of AI] mice had a significantly lower level of tumor-lytic ability at all NK cell-target cell ratios tested (p < O.OS-o.OOl) than did the splenocytes from CS7B1I6 mice (Fig. 2). Moreover, the NK cell frequency in blood of CS7B1I6 mice was significantly higher (p < O.OS) than that of AI] mice (Fig. 3). Table 2 illustrates the effect of in vivo administration of the powerful NK cell stimulant, ATRA, on the absolute numbers of NK cells. The results indicate no significant difference, in either the bone marrow or spleen, in the absolute numbers of NK cells between ATRA-treated and vehicle-administered mice within the same strain. By contrast, however, both the powerful NK stimulants, indomethacin and interleukin-2, significantly elevated the absolute numbers of
Table 1. Total cellularity and total numbers of lymphocytic cells in the spleen and bone marrow of A/] and C57B1I6 mice. Organ
Mouse strain
Total nucleated cells (xl 06) (mean ± s.e.)
Total lymphocytic cells (xl 06)" (mean ± s.e.)
Spleen
C57B1I6 AI]
110.3 ± 2.6 87.4 ± 2.8"-
86.9 ± 3.2 70.2 ± 5S:-
Bone marrow
C57B1I6 AI]
19.6 ± 0.5 15.0 ± OS'
5.2 ± 0.5 4.0 ± 0.8
Mean ± SE: 25-35 mice/strain. ':- p < 0.05 ':- Determined from differential counts of 1000 nucleated cells and converted via the known total cellularity to absolute numbers per spleen, or per femur.
Strain differences in NK cell immunity . 29 6
..
5
.! 'i
4
"'c
~
.."+::!
3
(/)
2
•
C57BII6
•
AlJ
u
0
~ ~
z
0
Spleen
Bone Marrow
Figure 1. The absolute numbers of NK(ASGM-l+) cells in the spleen and bone marrow (one femur) of C57B1I6 and A/] mice. The NK(ASGM-l+) cells in the spleen, but not the bone marrow of C57B1I6 mice are significantly more numerous (p < 0.05) than that of the spleen of A/] mice. Mean ± SE: 6-8 mice/group.
30..---------------------, --
C57BII6
- - - AlJ
0+---.------,,---...---........----.-----..---1 1 :3
1:6
1:12
1 :25
1 :50
1:100
Figure 2. Percent specific lysis of YAC-l lymphoma target cells by spleen cells from C57B1I6 and A/] mice. Mean ± SE: 5-7 mice/E:T ratio.
NK cells in both the spleen and the bone marrow within each strain, relative to vehicle-administered controls (Table 3). Figures 4 and 5 indicate the proportions of lymphocytic cells, in the spleens and bone marrow of both strains, whose immediate precursors had incorporated tritiated thymidine, a marker of DNA synthesis, at 1 h or 5 h earlier. At 1 h (Fig. 4) after a single injection of the radio-labelled DNA precursor, the proportions of lymphocytes bearing the isotope in the spleen and bone marrow, were similar between the two strains. Even at 5 h after isotope administration (Fig. 5), there
30 . A. L. WHYTE and S. C. MILLER Table 2. Effects of ATRA on the absolute numbers of NK(ASGM-1+) cells in the spleen and bone marrow of A/} and C57B1I6 mice. Organ
Mouse strain
Treatment
Absolute number of NK(ASGM-1+) cells (x1Q6)"
Spleen
C57B1I6
None b Corn oil vehiclec ATRAd
07 ± 0.44e 4.18 ± 0.22 3.74 ± 0.24
Bone marrow
C57B1I6
None Corn oil vehicle ATRA
0.45 ± 0.07 0.34 ± 0.04 0.29 ± 0.05
Spleen
A/}
None Corn oil vehicle ATRA
2.88 ± 0.32 2.40 ± 0.34 3.10 ± 0.26
Bone marrow
A/}
None Corn oil vehicle ATRA
0.36 ± 0.08 0.26 ± 0.03 0.29 ± 0.04
"Determined from differential counts of 1000 nucleated cells and converted via the known total nucleated cell count to absolute numbers per spleen, or per femur. bNo treatment. cCorn oil was given orally by gavage once/day for 5 days. dAll trans retinoic acid was administered (5 mg/kg) once/day for 5 days. eMean ± SE: 6-8 mice/group.
4
...... ~ ......
3
::IE CJ
2
...+•
rn c(
•
C57BV6
IIJJI AlJ
lie: Z
0-1-----
blood
Figure 3. NK(ASGM-t+) cells as a proportion of all lymphocytes in the blood of C57B1I6 and A/] mice, determined from differential analyses of 1000 total lymphocytes per hematologically stained cytospot. Mean ± SE: 6-8 mice/group. p < 0.05.
Strain differences in NK cell immunity .
31
Table 3. Effects of indomethacin or interleukin-2 on the total numbers of NK(ASGM-1+) cells in AI] and C57Bl/6 mice. Organ
Mouse strain
Treatment
Absolute numbers of NK(ASGM-1+) cells (x106)a
Spleen
C57B1I6
vehicle b indomethacin< interleukin-2d
5.4 ± O.4 e 7.2 ± 1.6 13.9 ± 3.1"
Bone marrow
C57B1I6
vehicle indomethacin interleukin-2
0.4 ± 0.04 1.0 ± 0.05':1.4 ± 0.2"-
Spleen
AI]
vehicle indomethacin interleukin-2
2.8 ± 0.1 3.9 ± 0.6" 6~4 ± 0.9"
Bone marrow
AI]
vehicle indomethacin interleukin-2
0.4 ± 0.03 0.9 ± 0.08'" 1.5 ± 0.4"
aDetermined from differential counts of 1,000 nucleated cells and converted via the known total nucleated cell count to absolute numbers per spleen or per femur. bThe vehicle for indomethacin is 1 ml absolute ethanol in the drinking water; the vehicle for interleukin-2 is 0.1 ml sterile phosphate buffered saline, containing 1% homologous serum, injected i.p. 3x/day for 4 days. <5 mg dissolved in 1 ml absolute ethanol, diluted in the drinking water to a final concentration of 10 jlg/ml. dRecombinant, human, given as bolus i.p. injections 3x/day for 4 days, each injection containing 12x106 units interleukin-2. eMean ± SE: 8 mice. "-p < 0.05-0.001; relative to vehicle-treated.
15 'C
.!! Gi
.c .!! CD
c
:c
E>.c
":"
12 •
C57BII6
fIrIII A1J
9 6
~
'#.
3
0
Spleen
Bone Marrow
Figure 4. Tritiated thymidine labelling index of lymphocytes in the spleen and bone marrow of C57Bl/6 and AI] mice. The in vivo time availability of the isotopically labelled, DNA precursor was 1 h - at which time the organs were sampled. Mean ± SE: 6-8 mice/group.
32 . A. L. WHYTE and S. C. MILLER Table 4. Effects of ATRA on the absolute numbers of 3H-thymidine-labelled lymphocytic cells in the spleen and bone marrow of AI] and C57B1I6 mice. Organ
Mouse strain
Treatment
Absolute numbers of H-thymidine labelled cells (x10 6).,b
Spleen
C57Bl/6 C57Bl/6
Corn oil vehiclec ATRAd Corn oil vehicle ATRA
6.74 ± 1.46< 5.57 ± 0.30 3.82 ± 0.50 3.90 ± 0.50
C57B1I6 C57Bl/6
Corn oil vehicle ATRA Corn oil vehicle ATRA
1.40 ± 0.14 1.45 ± 0.14 1.06 ± 0.19 0.89 ± 0.07
A/] AI] Bone Marrow
AI] AI]
'Determined from differential counts of 1000 nucleated cells and converted via the Known total nucleated cell count to absolute numbers per spleen, or per femur. bH-thymidine was given as a single pulse injection, i.p. (1 pCi/mg body wt), at the conclusion of 5 days of ATRA administration in the diet. cCorn oil was given orally by gavage once/day for 5 days. dAll trans retinoic acid was administered (5 mg/kg) once/day for 5 days.
was no significant difference in the proportions of isotope-bearing lymphocytes when comparing the bone marrow organs between the two strains, or, when comparing the spleens between the two strains. Within each strain, the proportion of labelled lymphocytes in the bone marrow at 5 h was approximately 3-fold that observed at 1 h.
15 'tI
.!! 'ii .c
12
"c
9
••
.!!
'6
C57BII6
AlJ
E
>-
i
~
6
..,
.,.
3 0
Spleen
Bone Marrow
Figure 5. Tritiated thymidine labelling index of lymphocytes in the spleen and bone marrow of C57B1I6 and A/] mice. The in vivo time availability of the isotopically labelled, DNA precursor was 5 h - at which time the organs were sampled. Mean ± SE: 6-8 mice/group.
Strain differences in NK cell immunity . 33
2r-------------------, •
II
CS7B1/6 AlJ
-
'",.. o
MEL·14
MAC·1
Figure 6. Absolute numbers of lymphocytes (enriched for NK cells) in the spleen which bear the surface molecule recognized by mAb MEL-14 or Mac-I. Mean ± SE: 5 mice/group.
At the 5 h interval, however (Fig. 5), there was a 52% reduction (p < 0.001) in the proportion of labelled lymphocytes in AI] spleen vs that of AI] bone marrow, whereas, by contrast, the proportions of labelled lymphocytes in the spleen and the bone marrow of C57Bl/6 mice were almost identical. Table 4 indicates that no significant difference, either strain-dependent or organ-dependent, could be found in the absolute numbers of labelled lymphocytes when mice were injected with ATRA or its vehicle. Finally, Figure 6 illustrates the absolute numbers of splenic lymphocytes, highly enriched for NK cells after Dynabead separation, which bore the homing receptor, gp90MEL-t4 or the integrin, Mac-t. The absolute numbers of these Dynabead-enriched lymphocytes, the majority of which are NK cells, which bore either of these surface molecules, differed either only marginally (gp90MEL-14), or insignificantly (MAC-t) between the two strains.
Discussion The low cellularity of the bone marrow, the primary production site of all the hemopoietic cells, of mice of the AI] strain relative to that of C57Bl/6 mice appears to reflect the previously demonstrated low levels of GM-CFC's in AI] bone marrow compared with C57Bl/6 mice (53). The consequence of this phenomenon would be, as the present data demonstrate, significantly lower numbers in AI] mice, of hemopoietic cells in the spleen, a major recipient organ of mature cells originating in the bone marrow. We have recently shown that, indeed, AI] mice have significantly lower absolute numbers of lymphoid, granulocytic and monocytic - but not erythroid - cells in their spleen andlor bone marrow than do C57Bl/6 mice (54).
34 . A. L. WHYTE and S. C. MILLER
It appears that not all of the NK(ASGM-1 +) cells produced in the bone marrow of the AI] mouse (whose numbers are similar to those of C57Bl!6] mouse bone marrow) are leaving the bone marrow, as reflected by the significantly lower proportions of these cells in AI] blood vs C57Bl!6 blood. The significantly lower NK(ASGM-1+) cell content of AI] spleen, would correspondingly reflect the low blood levels, the spleen being the primary destiny for virtually all bone marrow-derived, newly generated NK cells, and the blood, being the only route out of the bone marrow. The proportions of 3H-thymidine labelled lymphocytes were similar in the corresponding organs of the two strains at 1 h after giving a single in vivo injection of the radioisotope, indicating that proliferation of the lymphoid precursors, as a whole, was comparable in the two strains, in spite of significantly lower levels of NK cells (and other lymphoid cells) in the spleen. At 5 h after the single injection of labelled nucleotide (tritiated thymidine), C57Bl!6 mouse bone marrow and spleen (a major recipient of bone marrow-produced Band NK cells), contained equivalent proportions of isotope labelled lymphoid cells. The observation that there was no decrease, and in fact, some increase, in the numbers of lymphoid cells in the bone marrow of C57B1/6] mice at 5 h (relative to the 1 h value for that organ), simply indicates the mitotic activity of those cells which incorporated the isotope at 1 h. The 2 daughter cells, at 5 h now each with it'~ own complement of isotopic grains derived from the par'ent cell, have added to the recordable (countable) numbers of cells «labelled» at 5 h. Furthermore, free isotope available in vivo, may be picked up and utilized by any such cells going into DNA-synthesis (and subsequently mitosis). These lymphoid cells would also contribute to the numbers of labelled cells at 5 h. However, there appears in AI] mouse bone marrow to be, as well, a considerable retention of newly produced (isotope labelled) lymphoid cells, such that the splenic proportions of these cells are less than half those of the bone marrow of that strain at 5 h post thymidine injection. Significantly lower numbers of NK cells in AI] blood, the only exit route from the bone marrow, further support retention of at least some of these cells in that organ. These results collectively suggest a level of production of NK cells in AI] mice equivalent at least to that of C57Bl6 mice, but, apparently, one or more phenomena are taking place: (i) considerable postproduction intramyeloid NK cell deathlabortion is occurring, (ii) many NK cells are themselves intrinsically incapable of emigrating from the bone marrow, or, (iii) NK cells are being inhibited by the endogenous myeloid-infra-structure, from leaving that organ. A low level of bone marrow-to-spleen trafficking of newly formed lymphocytes, as a whole, in AI] mice, may account not only for the significantly lower NK cell frequency in AI] blood (vs C57Bl!J) (and consequently, the lower absolute numbers of NK(ASGM-1+) cells observed in the AI] spleen vs C57Bl!7 spleen), but also for the lower total numbers of lymphocytes in AI] spleen. By contrast, the bone marrow of both strains, contains similar numbers of lymphoid cells, a fact, which together with the kinetic studies involving tritiated thymidine uptake, does indeed reflect similar production levels of lymphoid cells by the bone marrow of both strains. Band NK lymphoid cells are bone marrow-generated, virtually exclusively.
Strain differences in NK cell immunity .
35
Although ATRA has been shown to increase «NK activity», i.e., lysis of tumor cells (27), it was not known whether this ATRA-stimulated increase in lytic function may result from an increase in absolute number of NK cells in the presence of ATRA. The present results indicate, however, that even after 5 days of exposure to ATRA (spanning, thus, at least 4 NK cell generation cycles (13», no increment of NK cell numbers could be found, in either AI] or C57Bl6 mouse organs. This finding strengthens previous observations which have demonstrated distinct ATRA-enhanced subcellular events such as increases in perforin levels and mRNA levels in killer cells (29). By contrast, two other potent NK-stimulating agents, indomethacin and interleukin-2, have resulted, in both organs (spleen and bone marrow), of both strains of mice (AI] and C57B1I6), in significantly elevated numbers of NK cells, indicating the influence of these agents at strain (genetically) independent stages of new cell genesis. The influence of these agents is consistent with our previous observations in other strains of mice (40-43). The homing receptor gp90MEL-14, and the integrin, MAC-1, are essential to transendothelial migration of lymphocytes and other blood-borne cells. DYNABEAD separated splenocytes, enriched for NK cells, demonstrated no difference between the 2 strains in their capacity to bear the homing receptor andlor the integrin. These observations suggest equivalent competence of NK(ASGM1+) cells in both strains with respect to localizing at, and binding to, endothelial cell walls such as would be encountered prior to extravasation from splenic vascular channels into the splenic parenchyma, the site of NK cell function. In conclusion, it appears that the long-established «low NK cell-mediated activity» of mice of the AI] strain (vs the «high NK cell-mediated activity» of the C57B1I6 strain) may have a single cause, i.e., lack of substantial exit (because of intra-myeloid abortion andlor inhibition of emigration) of newly produced, NK cell progeny from their bone marrow birth site. References 1. TIMONEN, T., A. RANKL, E. SAKSELA, and P. HAYRY. 1979. Human cell-mediated cytotoxicity against fetal fibroblasts. III. Morphological and functional characterization of the effector cells. Cell. Immunol. 48: 121. 2. KUMAGAI, K., K. ITOH, R. SUZUKI, S. HUNUMA, and F. SAITOH. 1982. Studies of murine large granular lymphocytes. I. Identification as effector cells in NK and K cytotoxicities. J. Immunol. 129: 388. 3. POLLACK, S. B., and C. ROSSE. 1987. The primary role of murine bone marrow in the production of NK cells: a cytokine study. J. Immunol. 139: 2149. 4. KIESSLING, R., E. KLEIN, H. PROSS, and H. WIGZELL. 1975. Natural killer cells in the mouse. II. Cytotoxic cells with specificity for mouse moloney leukemic cells: characteristics of the killer cell. Eur. J. Immunol. 5: 117. 5. BIRON, C. A., and R. M. WELSH. 1982. Activation and role of natural killer cells in virus infections. Med. Microbiol. Immunol. 170(3): 155. 6. TALMADGE, J. E., K. M. MEYERS, D. J. PRIEUR, and J. R. STARKEY. 1980. Role of NK cells in .tumor growth and metastasis in beige mice. Nature 284: 622. 7. HANNA, N., and R. C. BURTON. 1981. Definitive evidence that natural killer cells inhibit experimental tumor metastasis in vivo. J. Immunol. 127: 1754.
36 . A. 1. WHYTE and S. C. MILLER 8. JANEWAY, C. A. 1989. A primitive immune system. Nature 341: 108. 9. HALLER, O. R. KIESSLING, A. ORN, and H. WIGZELL. 1977. Generation of NK cells: An autonomous function of bone marrow. J. Exp. Med. 145: 1411. 10. HACKETT, J., M. BENNETT, G. C. Koo, and V. KUMAR. 1986. Origin and differentiation of natural killer cells. ImmunoI. Res. 5: 16. 11. KALLAND, T. 1986. Generation of natural killer cells from bone marrow precursors in vitro. ImmunoI. 57: 493. 12. POLLACK, S. B., M. Y. LEE, E. COHEN, J. 1. LOTTSFELDT, and C. ROSSE. 1989. Modulation of murine natural killer cells by a granulocytosis-inducing tumor. Cancer Res. 49: 174. 13. MILLER, S. C. 1982. Production and renewal of murine NK cells in the spleen and bone marrow. J. ImmunoI. 129: 2282. 14. GREENBERG, A. H., and M. GREEN. 1979. Non-adaptive rejection of small tumor inocula as a model of immune surveillance. Nature 264: 356. 15. ZOLLER, M., D. BELLGRAW, I. AxBERG, and H. WIGZELL. 1982. Natural killer cells do not belong to the recirculating lymphocyte population. Scand. J. ImmunoI. 15: 159. 16. Ross, G. D., and V. VETVICKA. 1993. CR3 (CDllb/CD18) a phagocyte and NK cell membrane receptor with multiple ligand specificities and functions. Clin. Exp. ImmunoI. 192: 181. 17. SOMERSALO, K., J. TARKKANEN, M. PATARROYA, and F. SAKSELA. 1992. Involvement of ~2 integrins in the migration of human natural killer cells. J. ImmunoI. 149(2): 590. 18. DIAMOND, M. S., D. E. STAUNTON, A. R. FOUGEEROLLES, S. A. STACKER, J. GARCIAAGUILAR, M. C. GIBBS, and T. A. SPRINGER. 1990. ICAM-1 (CD54): A counter-receptor for Mac-1 (CDllb/CD18). J. Cell BioI. 111: 3129. 19. SMITH, C. W., S. D. MARLIN, R. ROTHLEIN, C. TOMAN, and D. C. ANDERSON. 1989. Cooperative interactions of LFA-1 and MAC-1 with intercellular adhesion molecule-1 in facilitating adherence and transendothelial migration. J. Clin. Invest. 82: 2008. 20. DIAMOND, M. S., J. GARCIA-AGUILAR, J. K. BICKFORD, A. 1. CORBI, and T. A. SPRINGER. 1993. The I domain is a major recognition site on the leukocyte integrin Mac-1 (CDIIb/ CD18) for four distinct adhesion ligands. J. Cell. BioI. 120: 1031. 21. BUTCHER, E. C. 1991. Leukocyte-endothelial cell recognition: three (or more) steps to specificity and diversity. Cell 67: 1033. 22. MILLER, S. C. 1994. The development of natural killer (NK cells) from Thy-1(10)Lin-Sca-1+ stem cells: Acquisition by NK cells in vivo of the homing receptor MEL-14 and the integrin MAC-I. ImmunobioI. 190(4-5): 385. 23. KISHIMOTO, T. K., M. A. JUTILA, and E. C. BUTCHER. 1990. Identification of a human peripheral lymph node homing receptor: A rapidly down-regulated adhesion molecule. Proc. NatI. Acad. Sci. USA 87: 2244. 24. CHRISTOPHER, F. 1., I. DUSSAULT, and S. C. MILLER. 1991. Population dynamics of natural killer cells in the spleen and bone marrow of normal and leukemic mide during in vivo exposure to interleukin-2. ImmunobioI. 184: 37. 25. MILLER, S. c., F. 1. CHRISTOPER, and I. DUSSAULT. 1992. Population dynamics of natural killer cells and T lymphocytes in murine spleen and bone marrow during the development of erythroleukemia: The effect of indomethacin. Nat. Immun. 11: 78. 26. DUSSAULT, I., and S. C. MILLER. 1993. Stimulation of natural killer cell numbers but not function in leukemic infant mice: A system primed in infancy allows survival in adulthood. Nat. Immun. 12: 66. 27. GOETTSCH, W, Y. HATORI, and R. SHARMA. 1992. Adjuvant activity of all-trans-retinoic acid in C57Bl/6 mice. Immunopharma. 14(2): 143. 28. GOLDFARB, R. H., and R. B. HERBERMAN. 1981. Natural killer cell reactivity: regulatory interactions among phorbol ester, interferon, cholera toxin and retinoic acid. J. ImmunoI. 126(6): 2129. 29. LIN, T. H., and T. M. CHU. 1993. Enhancement of perforin by retinoic acid is mediated by protein kinase C. Biochem. And Biophys. Res. Comm. 191(3): 937, 30. FEGAN, c., BAILEy-WOOD, S. COLEMAN, S. A. PHILLIPS, 1. NEALE, T. Hoy, and J. A.
Strain differences in NK cell immunity . 37 WITI'AKER. 1995. All trans retinoic acid enhances human LAK activity. Eur. J. Haematol. 54: 95. 31. HERBERMAN, R. B., M. E. NUNN, and D. H. LARVIN. 1975. Natural cytotoxic reactivity of mouse lymphoid cells against syngeneic and allogeneic tumors. I. Distribution of reactivity and specificity. Int. J. Cancer 16.216. 32. HALLER, O. A., M. GIDLUNG, J. T. KURNICH, and H. WIGZELL. 1978. In vivo generation of mouse natural killer cells: role of spleen and thymus. Scand. J. Immunol. 8: 207. 33. KAMINISKY, S. G., I. NAKAMURA, and G. CUDKOWICZ. 1983. Selective defect of natural killer and killer cell activity against lymphomas in SJL mice: Low responsiveness to interferon inducers. J. Immunol. 130: 1980. 34. ITOH, K., R. SUZUKI, Y. UMEZU, et al. 1982. Studies of murine large granular lymphocytes. II. Tissue, strain and age distributions of LGL and LAL. J. Immunol. 129: 395. 35. CHEERS, c., and I. F. C. McKENZIE. 1978. Resistance and susceptibility of mice to bacterial infection: genetics of listeriosis. Infect. Immun. 19: 755. 36. PENNINGTON, J. E., and R. M. WILLIAMS. 1979. Influence of genetic factors on natural resistance of mice to Pseudomonas aeruginosa. J. Infect. Dis. 139: 396. 37. BROWN, I. N., A. A. GLYNN, and J. PLANT. 1982. Inbred mouse strain resistance to Myobacterium lepraemurium follows the Ity/Lsh pattern. Immunology 47: 149. 38. NESBITI', M. N., and E. SKAMENE. 1984. Recombinant inbred mouse strains derived from A/J and C57B1I6J: A tool for the study of genetic mechanisms in host resistance to infection and malignancy. J. Leuk. BioI. 36: 357. 39. YOSHIDA, A., Y. KOIDE, M. UCHIJIMA, and T. O. YOSHIDA. 1995. Dissection of strain difference in acquired protective immunity against Mycobacterium bovis Calmette-Guerin Bacillus (BCG). Macrophages regulate the susceptibility through cytokine network and the induction of nitric oxide synthetase. J. Immunol. 155: 2057. 40. ROSENSTEIN, M., S. E. ETTINGHAUSEN, and S. A. ROSENBERY. 1986. Extravasation of intravascular fluid mediated by the systemic administration of recombinant interleukin-2. J. Immunol. 137: 1735. 41. KOHLER, P. c.,J. A. HANK, K. H. MOORE, B. STORER, R. BECHHOFER, R. HONG, and P. M. SONDEL. 1989. Phase I clinical trial of recombinant interleukin-2: A comparison of bolus and continuous intravenous infusion. Cancer Invest. 7: 213. 42. WEST, W. H., K. W TAUER, J. R. YANNELLI, G. D. MARSHALL, D. W. ORR, G. B. THURMAN, and R. K. OLDHAM. 1987. Constant infusion recombinant interleukin-2 in adoptive immunotherapy of advanced cancer. N. Engl. J. Med. 316: 898. 43. GATELY, M. K., T. D. ANDERSON, and T. J. HAYES. 1988. Role of Asialo-GMI-positive lymphoid cells in mediating the toxic effects of recombinant IL-2 in mice. J. Immunol. 141: 189. 44. MILLER, S. c., and A. C. SHATZ. 1991. Relationship between large and small tumor-binding cells in the spleen and bone marrow. Nat. Immun. & Cell Growth Reg. 10: 320. 45. DUSSAULT, I., and S. C. MILLER. 1994. Decline in natural killer cell mediated immunosurveillance in aging mice - a consequence of reduced cell production and tumor-binding capacity. Mech. Ageing & Dev. 75: 115. 46. WAHL, L. M., and L. LAMPEL. 1987. Regulation of human peripheral blood monocyte collagenase by prostaglandins and anti-inflammatory drugs. Cell. Immunol. 105: 411. 47. GOLDMAN, R. 1984. Effect of retinoic acid on the proliferation and phagocytic capability of muring macrophage-like cell lines. J. Cell. Physiol. 120: 91. 48. KASAl, M., M. IWAMORI, #. NAGAI, K. OKUMURA, and I. TADA. 1980. A glycolipid on the surface of mouse natural killer cells. Eur. J. Immunol. 10: 175. 49. YOUNG, W W, S. HAKOMORI, J. M. DURDIK, and C. S. HENNEY. 1980. Identification of ganglio-N-temiosylceramide as a new surface marker for murine natural killer (NK) cells. J. Immunol. 124: 199. 50. STOUT, R. D., G. A. SCHWARTING, and J. SUTTLES. 1987. Evidence that expression of AsialoGm1 may be associated with cell activation. J. Immunol. 139: 2123. 51. BECK, B. N., S. GILLIS, and C. S. HENNEY. 1982. Display of the neutral glycolipid ganglio-
38 . A. L. WHYTE and S. C. MILLER N-tetraosylceramine (Asialo Gm1) on cells of the natural killer and T lineages. Transplantation 33: 118. 52. RODER, J. c., and R. KIESSLING. 1978. Target-effector interaction in the natural killer cell system. I. Covariance and genetic control of cytolytic and target-cell binding subpopulations in the mouse. Scand.]. Immunol. 8: 135. 53. POZZULO, G. N., E. SKAMENE, and F. GERVAIS. 1993. Bone marrow cell response following induction of acute inflammation in different strains of mice. Inflammation 17(6): 677. 54. MILLER, S. c., and S. L. KEARNEY. 1998. Effect of in vivo administration of trans retinoic acid on the hemopoietic cell populations of spleen and bone marrow: Profound strain differences between AI] and C57B1I6] mice. Lab. Animal Sci. 48 (1): 14. SANDRA C. MILLER, Ph.D., Department of Anatomy and Cell Biology, McGill University, 3640 University Ave, Room 2/28, Montreal, Quebec, Canada H3A 2B2