Low natural in vivo resistance to syngeneic leukaemias in natural killer-deficient mice

Immune characteristics of the beige-nude mouse. A model for studying immune surveillance

Double immunodeficient mice, homozygous for both the beige and the nude genes, were developed on C57BL/6N-beige and N:NIH(S)-nude background through five steps of mating. The animals were healthy and had a life span comparable to that of regular nude mice. Beige and regular nude mice showed a significant difference in natural killer (NK) activity, with mean +/- SE values of 6 +/- 1.0% and 25 +/- 2.8%, respectively. The response to the T-cell mitogens phytohaemagglutinin and concanavalin A was similar in beige and NIH mice of the same genotypes for the nude genes. Both groups of nu/nu mice also showed some response in these assays, as well as in the plaque-forming cell assay, indicating that nude mice are not completely devoid of functional T or T-like cells. The nude genes were found to increase the response to the B-cell mitogen lipopolysaccharide in both the beige and NIH animals, whereas the response was reduced in mice of the beige genotypes. The viable, low NK beige-nude mice here reported on may be a valuable tool for studying factors involved in host defence against tumours. Interestingly, no spontaneous tumours have so far been observed in such mice.

Adaptive immune defects and delayed rejection of allogeneic tumor cells in beige mice

The effect of the beige (bg) mutation on adaptive allogeneic tumor rejection was examined by monitoring tumor cell survival in vivo using [131I]iododeoxyuridine-prelabeled cells. Accelerated elimination of allogeneic tumor cells normally begins 8 days after ip injection and is due to active immune responses. Two independent mutations to beige on two different inbred backgrounds (C57BL/6J bgJ and DBA/2JCo bg8J) were tested, and bg/bg mice showed a 1-day delay in immune elimination of allogeneic cells. This delayed rejection was not due to a defect in clearing label from dead cells, nor to an inability to effect antibody-induced killing in vivo. Both humoral and cell-mediated responses against the allogeneic tumor cells were significantly lower in bg/bg than in +/bg mice.

Short-term and long-term effects of estrogen on lymphoid tissues and lymphoid cells with some remarks on the significance for carcinogenesis

Estrogens have long been thought to play a role in regulating the immune system. The difference in some types of immune responses between males and females is well-known, as is the pronounced thymic involution induced by exogenous estrogens. Estrogens stimulate some aspects of macrophage activity and, depending on dose and mitogen, inhibit or stimulate lymphocyte proliferative response in vitro. Another example is the estrogen effect on the delayed type hypersensitivity response. A broad review is given of such estrogen effects on lymphoid tissue and immune response. Most of the studies published so far are phenomenological. However, the recent description of estrogen receptors in the thymus and in some lymphocyte subpopulations, as well as a deeper understanding of regulating factors in the immune system, open the possibility of a more detailed understanding of the estrogen mechanism of interference. Estrogen effects in adults are reversible. After treating neonatal mice with the synthetic estrogen diethylstilbestrol (DES), disturbances are induced in lymphocyte populations and lymphocyte functions which are permanent and irreversible. Lymphocytes from adult, neonatally DES-treated female mice have a reduced mitogen response to ConA and LPS (T and B cell mitogen) and the delayed type hypersensitivity response is depressed. A detailed analysis demonstrated a decreased T helper cell population. The activity of Natural Killer cells is permanently reduced and this functional impairment is related to a decreased number of these cells, in turn determined at the bone marrow level. The same animals have an increased sensitivity to chemical carcinogens (methylcholanthrene) and they spontaneously develop epithelial changes in the uterine cervix which morphologically are similar to adenocarcinoma. The association between estrogen-associated malignancy and estrogen effects in lymphocyte functions deserves further study.

Influence of 12-O-tetradecanoylphorbol-13-acetate (TPA) on the susceptibility of K562 to natural cytotoxicity: evidence for clonal variation in differentiation-induced changes of lytic sensitivity

The effect of 12-O-tetradecanoylphorbol-13-acetate (TPA) on the sensitivity to NK cell-mediated lysis of two cloned populations of K562 which exhibit marked and stable differences in their susceptibility to natural cytotoxicity has been examined. Culture in medium supplemented with TPA concentrations of l ng/ml or more invariably caused a decrease in the susceptibility of the sensitive clone E10/P2, whereas treatment of the relatively resistant clone F9/P2 with TPA under identical conditions caused a significant increase in susceptibility to natural cytotoxicity. In both cases the change in susceptibility occurred within 1 day of culture in TPA and was rapidly reversible following removal of the inducing agent. The changes in resistance to natural cytotoxicity induced by TPA were independent of variations in osmotic fragility and were not attributable to alterations in NK cell binding capacity as determined by cold competition analysis. In contrast to the effect of TPA, exposure of E10/P2 and F9/P2 to interferon (IFN) caused a reduction in sensitivity to natural cytotoxicity of both populations which was associated with a decreased capacity to compete for lysis of labelled target cells. These data suggest that the effects of differentiating agents on target susceptibility to NK cell lysis are variable and that responses to TPA are clonally distributed within cell populations.

Prospective study of natural cytotoxicity in peripheral blood of patients with nonlymphoid solid malignancies

Natural killer (NK) cells have been implicated as an initial immunosurveillance mechanism for carcinogenesis in humans. Work in the murine system as well as the findings of depressed NK activity in patients with advanced malignancies and the discovery of increased incidences of cancer in humans congenitally deficient in NK ability have supported this. Few prospective studies have demonstrated a prognostic change in NK activity with respect to malignant disease course. In 32 healthy donors, NK activity against K562 was determined. No race or sex difference existed with respect to NK cell function. Esophageal (5), bronchogenic (3), breast (3), cervical (3), and endometrial (1) cancer patients who had received no prior chemotherapy were compared to controls. All patients subsequently received radiotherapy. Prior to such treatment NK activity could not be associated with stage of malignancy. Of the 15 patients studied, 11 were sequentially followed. Five of eight patients with stable or improving clinical courses as assessed by weight and Karnofsky scores were found to have increasing NK activity. Two of three patients with poor clinical courses presented with subnormal killing which never rose to normal while the third declined to subnormal before expiring. Esophageal, cervical, and endometrial carcinoma patients all presented with low or subnormal NK activity. Of these, only cervical and endometrial cancer patients exhibited an increase to normal levels.

Exposure of neonatal female mice to diethylstilbestrol persistently impairs NK activity through reduction of effector cells at the bone marrow level

Female NMRI mice neonatally treated with diethylstilbestrol (DES) show persistent impairment of various immune parameters, including NK activity. The present study demonstrated that NK activity was reduced in all lymphoid compartments and thus is not due to a simple redistribution of effector cells. Poly I:C was unable to augment NK activity in DES treated animals in vitro and in vivo. This defect was not attributable to inability to produce interferon in response to poly I:C since interferon-beta was also unsuccessful in augmenting NK activity in vitro. Moreover, the lack of response was not dependent on alterations in macrophage function, as evident from experiments showing that macrophages from DES-exposed animals resulted in a normal enhancement of NK activity in response to poly I:C when mixed with lymphocytes from control animals. Studies at the single cell level revealed that the reduced NK activity was the result of a reduced number of target-binding effector cells, and that the individual cells actually binding a target killed the target in an apparently normal fashion. Adoptive transfer of bone marrow cells between control and neonatally DES-treated animals showed that the reduced number of effector cells was determined at the bone marrow level.

The cooperative effect of the satin and beige mutations in the suppression of NK and CTL activities in mice

The functional activity of natural killer (NK) cells has been found to be modulated by several point mutations associated with coat color. The most commonly studied gene, beige (Bg), has been found to block a postrecognition event in the lytic cycle. Four other coat color mutations in the mouse (satin, leaden, fuzzy, pale ears) were studied for their effect on NK cell function, and only one, satin (Sa), was found to be suppressive. When both the Sa and Bg mutations were present in the same animal, their effects were synergistic in the suppression of NK levels. Normal numbers of NK cells were present in these double mutants, as determined by the frequency of IgG2b binding cells and by antiasialo GM1 staining. The ability of Sa/Bg NK cells to recognize and bind targets suggests that the defect is localized in the postbinding cytolytic pathway. These genes were not specific for NK cells and also suppressed alloimmune cytolytic T lymphocyte function. Since Sa/Bg mice are much more suppressed in NK function than Bg mice, we suggest that this double mutant may be a better model for NK deficiency in vivo.

Dichloromethylene diphosphonate (Cl2MDP) reduces natural killer (NK) cell activity in mice

Daily administration of dichloromethylene diphosphonate (Cl2MDP) to (C57BL/6 X DBA/2) F1 hybrid mice, from two days of age (10 mg of P/kg body weight), resulted in a marked impairment of natural killer (NK) activity of spleen cells against YAC-1 lymphoma cells. The suppressive effect increased with the duration of the treatment. Cessation of the treatment led to a rapid recovery (in 2 weeks) of NK activity while the osteopetrotic bone lesions persisted. Thus, the loss of natural killing cannot be explained by the simple reduction of bone marrow volume secondary to Cl2MDP-induced osteopetrosis. However, as NK cells are considered to be dependent on the bone marrow because they cannot be sustained by extramedullary production, a direct effect of Cl2MDP on the generation of NK cell precursors by the bone marrow was not excluded. Cl2MDP was not directly toxic to the fully differentiated splenic NK cells, since the addition of Cl2MDP to the in vitro assay (10(-5)-10 micrograms/ml) did not reduce cytotoxicity. These studies suggest that impairment of NK activity during Cl2MDP treatment may have clinical toxicologic implications since NK cells have been suggested to play an important role in natural host defenses against infection and neoplasia.

Analysis of metastatic spread and growth of tumor cells in mice with depressed natural killer activity by anti-asialo GM1 antibody or anticancer agents

The mechanism of artificial and spontaneous metastases of tumor was analyzed in B16 melanoma cells and C57BL/6 mice by using anti-asialo GM1 antibody and anticancer agents. Single administrations of 500 micrograms anti-asialo GM1 antibody resulted in significantly decreased NK activity in spleen cells of C57BL/6 mice, lasting 10 days from the day following administration. Treatment with anti-asialo GM1 antibody never decreased the function of T lymphocytes measured by blastogenesis with phytohemagglutinin or T cell growth factor. The tumoricidal functions of activated macrophages but not of resident macrophages were decreased by in vivo treatment with anti-asialo GM1 antibody. The anti-asialo GM1 antibody was evaluated in terms of the enhancing effect on pulmonary metastases with regard to the timing of administration. Treatment with anti-asialo GM1 antibody 1 day before or on the day of tumor inoculation resulted in a substantial increase in the number of artificial pulmonary metastases. In the experimental system of spontaneous metastases, anti-asialo GM1 antibody most effectively increased the number of pulmonary metastases when administered 1-2 weeks before the removal of primary tumor, when the tumor cells are thought to be released into blood circulation from the primary site. In addition, accelerated growth of transplanted tumors at the primary site was observed in mice treated with anti-asialo GM1 antibody. These results strongly suggest that anti-asialo GM1 antibody enhances the incidence of in vivo tumor metastases and the growth of transplanted tumor mainly by suppressing the function of NK cells. The maximum effective dose (MED) of mitomycin C or its derivative (M-83) suppressed NK activity significantly, and pretreatment with these anticancer agents enhanced the growth of the artificial pulmonary and liver metastases. In contrast, the MED of cDDP showed no effect on the NK activity or the numbers of pulmonary and liver metastases. These results indicate that the depression of NK activity induced by chemotherapy results in the promotion of metastatic disease. From these studies it can be concluded that NK cells have a key role in the control of metastases of malignant disease, and that support of NK activity is very important for the prevention of metastases.

Gamma-interferon (IFN-gamma) produced during effector and target interactions renders target cells less susceptible to NK-cell-mediated lysis

Human mitomycin C-treated PBL were mixed with cells of an NK sensitive hybrid cell line (PUTKO-I). A fraction of tumor cells survived this treatment and could be recovered from the cultures. These surviving cells were completely NK-resistant and this property persisted for 2-3 weeks after cultivation in fresh medium. Treatment of a clone (C13) of PUTKO-I with PBL-PUTKO mixed lymphocyte-tumor-cell culture (MLTC) supernatants resulted in a marked reduction in NK sensitivity after 8-12 h of treatment. The kinetics of induction of NK resistance by MLTC supernatants was similar to that of purified IFN-gamma and was faster than for IFN-alpha. The active component in the supernatants was characterized as a mixture of IFN-gamma and IFN-alpha based on neutralization of activity with specific antisera. The role of mycoplasma contamination was investigated and it was found that cell lines free of detectable mycoplasma stimulated production of NK-protective activity by PBL and this activity was neutralized by anti-IFN-gamma serum. Separation of PBL on discontinuous Percoll gradients demonstrated a correlation between the NK activity of cell fractions and their ability to produce IFN in response to tumor cells. Taken together, the selection-dependent variations in NK sensitivity, the kinetics of IFN production and induction of resistance suggest that tumor cells may be able to escape elimination by NK cells due to protection by IFN produced by the effector-cell-containing population.

Radiofrequency destruction of the tuberoinfundibular region of hypothalamus permanently abrogates NK cell activity in mice

Natural killer (NK) cells have an important role in non-adaptative resistance to tumours and their metastatic spread in vivo. Maturation of NK cells and the intensity of their activity are affected by many endogenous and external factors, as well as by regulatory cells. The possibility that some effects of the central nervous system on tumour resistance are mediated via NK activity has also been suggested. Destruction of the tuberoinfundibular region of the hypothalamus in rodents led to a significant increase in tumour growth. We show here that destruction of its ventromedial, dorsomedial and arcuate nuclei persistently abrogates NK activity in mice. By contrast, cortical lesion and operative stress depress it partially, and for a brief period only. Abrogation is the result of a block of NK lineage maturation, causing a severe decrease in the number of large granular lymphocytes (LGL), a lymphocyte population associated with NK activity. Macrophage, B- and T- lymphocyte functions, however, are not significantly affected. Agents inducing NK-cell maturation or activation such as polyinosinic-polycytidylic acid (poly(I:C], interferon (IFN) and interleukin-2 (IL-2) restore NK activity, and normalize the number of LGL.

Human breast cancer and impaired NK cell function

Recent advances in tumor immunology have led to the discovery of a new lymphoid cell with unique antitumor activity. Natural killer (NK) cells form an antitumor surveillance system and appear to be vital in preventing tumor growth and metastasis in animal models. We studied NK activity in patients with benign and malignant breast disease, using a chromium-51 release microtiter cytotoxicity assay with K562 cells as targets. Compared with benign controls, patients with malignancies had significantly depressed NK-mediated lysis (P less than 0.01). Furthermore, lysis in those with advanced disease (stages II, III, and IV) was significantly less than in those with limited disease (stage I) (P less than 0.01). NK activity was not correlated to estrogen or progesterone receptor states. Positive correlation of a depressed natural killer activity with the extent of tumor spread supports the concept of an NK cell immune surveillance system in breast cancer and emphasizes its importance in this malignancy.

Natural killer cell activity in inflammatory joint disease

The natural killer (NK) cell activity of unfractionated peripheral blood and synovial fluid mononuclear cells from patients with inflammatory joint disease was measured in a short-term assay using the human tumour cell line, K562, as the target. The mean values for peripheral blood NK activity of the various groups (controls, rheumatoid arthritis (RA), ankylosing spondylitis (AS) and psoriatic arthritis (PsA] were similar although the rheumatoid group showed the widest range. However, the NK activity of PsA patients (23.74 +/- 10.14) was significantly lower than that of the controls (31.63 +/- 10.8, 0.05 greater than P greater than 0.01). Almost without exception, NK activity was found to be considerably lower in synovial fluid than in paired blood samples (p less than 0.01).

A modified short-term cytotoxicity test: assessment of natural cell-mediated cytotoxicity in whole blood

Using whole blood from normal subjects, we have observed natural killing of K562 cells in a 4 h 51Cr-release assay comparable with that shown by separated PBMC and whole blood depleted of serum components. Separated plasma was not toxic towards K562 targets, and failed to potentiate the level of PBMC cytotoxicity through ADCC. The presence of red blood cells did not influence natural killing. The natural cytotoxicity of whole blood was augmented by interferon and depressed by prostaglandins E1 and E2. Studies with appropriate control blood fractions show that cytotoxicity tests with whole blood provide results reflecting natural cell-mediated cytotoxicity.

Depressed levels of granular lymphocytes with natural killer (NK) cell function in 247 cancer patients

The HNK-1 (Leu-7) monoclonal antibody was used to enumerate and characterize the level of blood granular lymphocytes in 247 cancer patients. The results were compared to 146 control individuals. A fluorescence-activated cell sorter was used to purify blood HNK-1+ cells from cancer patients. The monoclonal antibody identified a homogeneous population of granular lymphocytes with greater than 95% purity. Conversely, virtually 100% of HNK-1- cells from cancer patients were agranular lymphocytes. These results were the same as previously observed in normal individuals, where the HNK-1+ cell fraction contained all the lymphocytes with spontaneous cytotoxicity in natural killer (NK) and killer (K) cell assays. The level of HNK-1+ cells in cancer patients correlated significantly with the patient's age and sex, with older individuals having higher levels and male patients containing a higher proportion than female patients. The levels in the cancer patients were significantly lower than normal controls (p = 0.04). When the results were subdivided by the histologic type of cancer, additional differences were noted. Compared to age and sex-matched controls, significantly depressed levels of HNK-1+ granular lymphocytes were observed in 49 patients with colon cancer (9.7% vs. 15.8%, p = 0.0001), 18 patients with lung carcinoma (11.7% vs. 27.0%, p = 0.0001), 24 patients with breast carcinoma (12.0% vs. 15.5%, p = 0.04) and 64 patients with head and neck carcinoma (15.9% vs. 19.1%, p = 0.05). However, there were no significant differences overall in the average HNK-1+ cell level of 66 patients with melanoma (13.0% vs. 13.5%, p = 0.75) and nine patients with sarcomas (15.8% vs. 14.3%, p = 0.71). Thus, this important subpopulation of granular lymphocytes with NK and K cell function was significantly depressed in most cancer patients. Accounting for the patient's age and sex and the histologic type of cancer was critical to interpreting the results.

Studies of the thymus in mice bearing the Lewis lung carcinoma. II. Modulation of thymic natural killer activity by thymulin (FTS-Zn) and the antimetastatic effect of zinc

Modulation of thymic natural killer (NK) cell activity by the serum thymic factor (thymulin or FTS-Zn) was demonstrated in mice carrying the Lewis lung carcinoma (3LL). Thymulin decreased the increased NK activity found in amputated (tumor) tumor-bearing mice and induced a significant NK activity in nonamputated mice, normally devoid of such activity. Zinc had no effect in either case. Histological studies revealed that zinc prevented the tumor-induced thymic atrophy. A clear-cut decrease in the number of lung metastases was observed in zinc-treated mice. This effect was suppressed by concomitant administration of thymulin to amputated mice. The thymic dependency of the anti-metastatic action of zinc as well as the mode of action of thymulin on NK cell activity in tumor-bearing animals is discussed.

Modulation of K562 cells with sodium butyrate. Association of impaired NK susceptibility with sialic acid and analysis of other parameters

Neuraminidase treatment of parental and butyrate-induced K562 tumor cells was associated with an increase in natural killer (NK) susceptibility of these target cells. The degree of enhancement with neuraminidase was significantly greater for the NK-resistant (NRR) butyrate-differentiated K562 cells so that the relative difference between the parental NK-sensitive (NKS) K562 line and its induced NKR variants, in terms of NK sensitivity, was no longer five- or six-fold but only two-fold. The predominant reason for the altered NK susceptibilities of the target cells after neuraminidase treatment was an increase in the target-cell-binding ability of these cells as assessed by a direct conjugate-forming cell assay using Percoll-enriched NK cells and cold target competition assays. The enhancement did not appear to be due simply to an increased membrane-membrane attraction caused by a reduction of net negative cell surface charges since protamine sulphate, a positively charged molecule, had no effect on NK activity. Compared with the NKS parental K562 tumor cells, the NKR butyrate-induced cells had 3.6- to 4.0-fold higher sialo-transferase activities and were associated with significantly greater amounts of cell surface sialic acid detected both in sialyl glycoproteins (2.2- to 2.9-fold higher) and particularly within ganglioside extracts (6.2- to 13.6-fold higher). In conformity with the marked neuraminidase enhancement of NK-mediated cytolysis of the butyrate-induced targets, these NKR cells were associated with significantly enhanced levels of neuraminidase-accessible sialic acid compared to the NKS parental K562 cell line. Other parameters such as sensitivity to superoxide radicals, intrinsic superoxide dismutase levels, altered membrane repair mechanisms and transferrin competition, were not significantly different between the NKS and NKR target phenotypes. Sugar inhibition studies demonstrated an enhanced inhibition against the butyrate-induced cells with a variety of neutral sugars. The degree of inhibition with phosphorylated sugars was comparable between the parental and induced K562 tumor target cells and is consistent with our previous findings showing that these hexose phosphates may be inhibiting cytolysis at a step independent of target-cell recognition.

Natural killer cell activity during premedication, anaesthesia and surgery

Natural killer (NK) cell activity of peripheral blood mononuclear cells was measured against K-562 target cells in a 51Cr release assay in eight patients undergoing total hip replacement surgery. Eight consecutive blood samples were taken from each patient. A significant increase of NK cell activity was observed after premedication with diazepam per os. The activity increased further during a combined anaesthesia (thiopentone + N2O + O2 + buprenorphene + pancuronium) and remained increased during surgery. Postoperatively, NK cell activity fell and remained depressed for a period of at least 5 days. The findings of this study indicate that premedication, anaesthesia and surgery cause a rapid and transient increase in NK cell activity, followed by a decline in activity postoperatively. The transient increase in activity may be explained by mobilization of natural killer cells from extravasal space, spleen or lymph nodes into the circulation. The clinical significance of the alterations in NK cell activity is unknown and needs further investigation.

Elevated antibody titers to Epstein-Barr virus and low natural killer cell activity in patients with Chediak-Higashi syndrome

Four Venezuelan patients with the autosomal recessive Chediak-Higashi syndrome (CHS) were studied. The results confirm the severe reduction in natural killer (NK) cell activity, as previously described and showed also a decline in the activity of cells involved in antibody-dependent cellular cytotoxicity (ADCC). No defect was found in the production of immunoglobulins and of specific antibodies to measles, varicella, herpes simplex, and cytomegalo viruses. Two of the patients had extremely high antibody titers to the Epstein-Barr virus (EBV) specific viral capsid antigen (VCA), to the restricted (R) component of the EBV-induced early antigen complex, and to the EBV-associated nuclear antigen (EBNA). These two patients had enlarged livers, spleens, and lymph nodes indicative of the lymphoproliferative phase. The other two patients were initially negative for all EBV-associated antibodies but seroconverted subsequently and, in the course of a year, also developed high antibody titers to VCA and R. In one of these patients the primary infection was accompanied by moderate signs of infectious mononucleosis (IM) followed after more than 6 months by persistent hepatosplenomegaly. The other patient also developed signs of a lymphoproliferative syndrome with hepatosplenomegaly and jaundice and died 8 months later. Such high anti-R titers are seen frequently in Burkitt's lymphoma, but rarely in other conditions. It is likely that the high antibody titers reflect an increased production of VCA and R due to defective NK and ADCC cell activities so that productively infected B lymphocytes are no longer eliminated before they have synthesized maximal amounts of antigens. The high anti-EBNA titers suggest normal T lymphocyte function. The possibility that the accelerated, lymphoma-like phase of the CHS involves EBV-transformed cells is discussed.

Solid tumor-derived target cell susceptibility to macrophages and natural killer/natural cytotoxic cells in the rat

Cytotoxic capacity of rat macrophages (M phi) and natural killer (NK)/natural cytotoxic cells (NC) against adherent growing, solid tumor-derived target cells was evaluated, modulating the activation status of effector cells and growth conditions of target cells. Testing a panel of target cells, cytotoxicity of NK/NC and M phi was strikingly correlated so that besides of target-cell binding structures basic lysability seems to be of influence with respect to cytotoxicity rates. Varying the in vivo growth conditions of target cells altered their lysability by M phi and NK/NC cells in the sense that ascitic versus subcutaneously (sc) grown tumors were more resistant to lysis. On the other hand, in vitro culturing did not influence susceptibility for M phi, but with some tumor lines increased lysis by NK/NC cells was observed. In the rat, the activation status of M phi and NC was not age-dependent, and NK cell activity only declined slowly with age. But cytotoxic potential of M phi obviously presents a strain characteristic, different from NK/NC cell activity, only the latter two correlating in different rat strains. Experiments to augment natural cytotoxic capacity revealed that application of Corynebacterium parvum (CP) activated M phi and NK/NC cells, while sc tumor implantation only resulted in increased NK/NC cell cytotoxicity, leaving M phi unaltered.

Thymosin alpha 1 restores NK-cell activity and prevents tumor progression in mice immunosuppressed by cytostatics or X-rays

The effect of thymosin against tumor progression was examined in mice immunosuppressed by cytostatics or X-ray irradiation. When pretreated with cytostatic agents, such as 5-fluorouracil (5-FU) or BCNU, or by X-ray, and then inoculated with P388 or L1210 leukemias, mice died rapidly within a few days. In these systems, thymosin alpha 1 given concomitantly with the cytostatic agents or after X-irradiation prevented rapid death and extended survival, although the mice eventually died with leukemia like normal mice inoculated with cells of the same tumor. Rapid death in the 5-FU-treated mice was also prevented by adoptive transfer of spleen cells from the donor mice if these had been treated with 5-FU plus thymosin alpha 1, but not if they had received 5-FU alone. However, the restorative activity of the donor spleen cells was abrogated by treatment with anti-asialo GM1, but not by treatment with anti-Thy 1 or anti-mouse Ig serum, suggesting that the effector cells in the spleen are NK cells. In fact, thymosin alpha 1, when given concomitantly with 5-FU or after X-irradiation, maintained the NK activity of spleen, which was damaged by treatment with 5-FU or X-irradiation alone. The present study indicates that thymosin alpha 1 exerts a preventive activity against progression of leukemias at least in part through an effect on NK cells or their progenitor cells.

Assessment of in vivo natural antitumor resistance and lymphocyte. Migration in mice: comparison of 125I-iododeoxyuridine with 111indium-oxine and 51chromium as cell labels

Clearance of IV-injected tumor cells has been correlated with levels of natural killer (NK) cell activity in recipient animals. Studies of in vivo tumor cell clearance strongly suggest a relationship between levels of NK cell activity and antitumor or antimetastatic effector function. This study outlines the applicability of three radiolabels, [125I]iododeoxyuridine, ( [125I]dUrd), indium-111-oxine chelate ( [111In]Ox), and chromium-51 (51Cr), to studies of tumor cell clearance in vivo. The suitability of these labels for analysis of the in vivo migration patterns of normal lymphocytes or thymus-derived T cells cultivated in vitro (CTC) is also discussed. The results indicate that [111In]Ox and 51Cr compare favorably with the more widely used [125]dUrd as radiolabels for the assessment of IV-injected tumor cell clearance from the lungs of mice. The rates of clearance of both [111In]Ox and 51Cr, like that for [125I]dUrd, correlate closely with levels of NK-cell activity of the host. Further studies with [111In]Ox reveal that treatment of recipients with anti-asialo GM1 serum, a regimen known to suppress NK-cell activity, demonstrates the appropriate reduction in isotope clearance from the lungs after NK suppression. However, clearance data obtained by monitoring levels of radioactivity in the liver after IV injection must be viewed cautiously, since the same cells labeled with [111In]Ox and [125I]dUrd had a different pattern of clearance from the liver. The same inconsistencies in clearance were observed when [111In]Ox and [125I]dUrd were injected intrafootpad (i.f.p.). Similar effects were observed when [111In]Ox or 51Cr was applied to studies of CTC migration. Levels of [111In]Ox and 51Cr remained high in the liver after IV injection, while [125I]dUrd was rapidly cleared. Normal spleen or thymic lymphocytes exhibited the expected homing to the spleen after labeling with [111In]Ox, indicating a suitability of this label for migration studies, except possibly in the liver. These results with CTC and normal lymphocytes should be considered during the formulation of immunotherapy protocols based on cell migration data, since the choice of radiolabel can result in widely divergent levels of radioactivity accumulated in some organs, and may not provide an accurate representation of the presence of viable, intact, or functional cells.

Thymus independence of hybrid resistance against a panel of T-cell lymphomas of H-2b origin

Adult thymectomy, followed by whole-body irradiation and reconstitution with fetal liver, was performed to study the T-cell dependence of F1 hybrid resistance to a panel of lymphomas of H-2b origin. Previously, the pattern of hybrid resistance against the same lymphomas was found to correlate with the pattern of NK-activity in a spectrum of F1 hybrids (Kiessling et al., 1975). We now show that hybrid resistance against three lymphomas of C57BL/6 origin, P-52-127-166, RBL-5 and EL-4 and against YLD, of C57L origin, is expressed in the absence of thymus. In another series of experiments, the effectors responsible for hybrid resistance to the transplanted lymphoma EL-4 were studied by reconstituting thymectomized and non-thymectomized C57BL mice with syngeneic bone marrow from NK-deficient beige mutant or wild-type C57BL donors. While the recipients of beige bone marrow had a clearly reduced tumor resistance, thymectomy did not decrease resistance further. This study supports the hypothesis that resistance to these lymphomas in F1 hybrids as well as in syngeneic mice is mainly mediated by natural killer cells.

Depressed spontaneous natural killing and interferon augmentation in patients with malignant lymphoma

A substantial proportion (44%) of peripheral blood lymphocyte samples from 41 patients with malignant lymphoma have been shown to have depressed or undetectable levels of natural cytotoxicity against the leukaemic cell line K562 in a 4-hr [51Cr]-release assay. No correlation was found between low levels of natural killer (NK) cell activity and either the age of the patients, total or differential white blood counts, or the type or stage of disease. Furthermore, pre-treatment of lymphocytes with human lymphoblastoid (Namalva) interferon failed to enhance NK levels in 5/11 patients with Hodgkin's disease and 5/8 patients with non-Hodgkin's lymphoma, and was in contrast to the response of control peripheral blood lymphocytes assayed under the same test conditions. The lack of responsiveness to interferon of peripheral blood NK cells from lymphoma patients was not wholly associated with those patients shown to have low levels of spontaneous NK activity.

The biology of the human natural killer cell

Natural killer (NK) cells in the human are a population of large granular lymphocytes (LGL) with at least one unique surface antigen not expressed on cells of other lineages. NK-target-cell interaction appears to involve carbohydrate recognition and, following binding, the NK cells are induced to generate O2-, transmethylate membrane phospholipids, and activate phospholipase A2. Some or all of these activities trigger a cascade of events which ultimately leads to the secretion of a substance toxic to the target cell. A variety of genes controls various steps in this cytolytic pathway. There is a good deal of evidence in the mouse, and some in the human, that NK cells play a role in host surveillance against tumor development, resistance to viral infections, and, possibly, hematopoietic regulation.

Resistance to infections in mice with defects in the activities of mononuclear phagocytes and natural killer cells: effects of immunomodulators in beige mice and 89Sr-treated mice

Beige mice, which are a homolog of the Chediak-Higashi syndrome, and mice treated with 89Sr to destroy the bone marrow provide animal models of defects in mononuclear phagocyte and natural killer cell functions. The innate resistance of these mice to viruses such as herpes simplex and encephalomyocarditis viruses, however, is normal. Moreover, treatment of the mice with immunomodulators such as Propionibacterium acnes (formerly designated Corynebacterium parvum) and pyran produced a significant increase in resistance to encephalomyocarditis virus. The antiviral effect of P. acnes in 89Sr-treated mice was exhibited during marked monocytopenia and without evidence for an inflammatory influx of macrophages into the peritoneal cavity. Treatment with P. acnes was also effective in increasing the resistance of beige mice to infection with Listeria monocytogenes. Thus, immunomodulators can be effective in mice that exhibit impaired macrophage and natural killer cell functions.

Spontaneous lymphocyte-mediated (NK cell) cytotoxicity in systemic sclerosis: a comparison with antibody-dependent lymphocyte (K cell) cytotoxicity

Spontaneous (NK cell) and antibody-dependent (K cell) cytotoxicity were investigated in 39 patients with systemic sclerosis (SS) and compared with that found in 52 normal controls. Cr-labelled Chang liver cells were used as targets in assays utilising both whole blood (WB) and peripheral blood mononuclear cells (PBM) as effectors. Patients with SS, who were severely affected by extensive visceral disease, were found to have significant impairment of both NK (p less than 0.005; p less than 0.05) and K (p less than 0.001; p less than 0.05) cell cytotoxicity by both effector systems, when compared with normal controls. These findings, which seem to be part of a wider defect in cell-mediated immunity, may provide a possible explanation for the described association of malignancy with systemic sclerosis.

Selective depletion of NK cell activity in vivo and its effect on the growth of NK-sensitive and NK-resistant tumor cell variants

Intravenous injection of rabbit anti-asialo-GM1 serum, an antiserum previouslY shown to eliminate splenic natural killer (NK) activity in vitro, profoundly depressed NK activity in CBA, DBA/2 and BALB/c nu/nu mice. The effect on NK activity was selective, as treatment of mice with anti-asialo-GM1 serum did not affect the development of other cytotoxic cells including cytotoxic macrophages following injection of poly I:C, or cytotoxic T cells in response to allogeneic cells. The role of NK cells in controlling tumor cell growth was investigated using an NK-sensitive (cl 27v-1C2) and an NK-resistant (cl 27av) subline of the murine lymphoma L5178Y. Initial studies showed that cl 27v-1C2 cells were at least 100 times less tumorigenic than were cl 27av cells in both syngeneic DBA/2 mice and BALB/c nu/nu mice. In addition, treatment of DBA/2 mice with poly I:C, which boosted NK activity, markedly depressed the growth of cl 27v-1C2 cells, but not of cl 27av cells. On the other hand, treatment of DBA/2 mice and BALB/c nu/nu mice with anti-asialo-GM1 serum led to a marked increase in tumorigenicity of cl 27v 1C2 cells, but had no effect on the tumorigenicity of cl 27av cells. In addition, the protection against cl 27v-1C2 growth afforded by poly-I:C treatment was abrogated by injection oif anti-asialo-GM1 serum. The possibility that the effects observed were caused by binding of the injected antibodies to the tumor cells was minimized by: (1) using a clone of tumor cells (cl 27v-1C2) that lacks chemically detectable asialo-GM1, and (2) pretreating animals with anti-asialo-GM1 rather than administering antiserum and tumor cells concurrently. These studies provided compelling evidence that NK cells could play an active role in controlling tumor growth. Selective depletion of NK activity by injection of anti-asialo-GM1 serum is a method which would be generally applicable to studying the role of NK cells in disease processes.

Evidence that hamster fibroblasts tumors emerge in nude mice through the process of two in vivo selections leading to growth factor "relaxation" and to immune resistance

The Chinese hamster lung fibroblast cell line (CC139), anchorage- and highly serum-dependent for growth is tumorigenic in nude mice. Tumors arise after 4 to 8 weeks following the inoculation of 5 X 10(5) cells. We have shown that all the emerging tumoral clones (more than 20 analyzed) have lost the growth factor dependence of the parental cells (Pérez-Rodriguez et al., 1981 a). To mimic this selection which occurred in vivo, we selected in vitro growth-factor-independent variants. These variants, GFI304 and GFI461, can proliferate in a serum-free medium supplemented with transferrin alone. This character is stable since it is not lost after the GFI variants have been cultivated in non-selective medium (serum) for more than 20 generations. In spite of their "autonomous" growth and anchorage independence. GFI variants are poorly tumorigenic. The observation of nodule formation and subsequent regression in nude mice, immunosuppressed (irradiation or cyclophosphamide) or not, led us to the conclusion that at least two selections in vivo are required for the tumoral expression of CC139 cells. One leads to a loss of growth factor requirement, the second towards a resistance to the immune surveillance mechanisms of the nude mice. The 4- to 8-week lag period of tumor formation may be accounted for by the spontaneous emergence in vivo of the two new characters necessary to bypass host growth restraints. This report supports the concept of stepwise progression and clonal evolution of preneoplastic cells in vivo and also indicates that tumorigenicity tests in nude mice should be interpreted with caution.

Human natural killer-like activity against mouse spleen cells

Unstimulated human peripheral lymphocytes were cytotoxic for normal mouse spleen cells and suppressed the in vitro antibody plaque-forming cell (PFC) response of these cells to sheep red blood cells and dinitrophenylated Ficoll. The cells in the lymphocyte population that were responsible for the immunosuppression had properties of natural killer (NK) or NK-like cells in that they were: (a) non-E-rosetting, (b) nonadherent, (c) unaffected by treatment with anti-human immunoglobulin plus complement, (d) cytotoxic against an established human NK target, K562 leukemic cells, and (e) partially inactivated by mitomycin C. Addition of the human NK-like cells to mouse spleen cell cultures at the time of antigen addition and at an effector cell to target cell ratio as low as 0.67:1 resulted in 85 to 96% suppression of the PFC response. Addition of NK-like cells to cultures 18 h before harvesting in 5-day cultures required higher concentrations and ratios (2.7:1) of effector to target cells to significantly suppress the PFC response. The data suggest that human NK-like activity in suppression of the mouse PFC response is due to killing of the targets. The mouse spleen cell PFC system represents a potential model for assessment of human NK activity that is quite dramatic in its effect and can be used in addition to the well known 51Cr-release assay. Also, since the mouse spleen cell is a normal cell, it provides a model in the PFC system for studying the mechanism of NK regulation of normal cellular function. An additional finding of this study was the observation that E-rosetting T cells significantly enhanced the mouse PFC response. Thus, human peripheral lymphocytes contain discrete cellular population that either enhance or suppress the mouse PFC response.

Decreased sensitivity of immunoselected and NK-selected YAC lymphoma sublines to hybrid resistance in vivo

Two sublines of the Moloney lymphoma YAC, selected by alternating in vitro exposure to anti-MCSA + complement and in vivo passage in preimmunized hosts, had a decreased or undetectable expression of MCSA. These 'immunoselected' sublines were compared with the original YAC line with regard to their ability to grow in a panel of semisyngeneic F1 hybrids. Natural hybrid resistance to YAC, previously found to be mediated by NK cells, affected the immunoselected sublines to a much smaller extent. This was further corroborated by the fact that the same sublines showed a decreased sensitivity to the in vitro lysis by NK cells from the same hybrid genotypes. Another set of YAC variants were produced by repeated in vitro exposure to NK cells and intermittent passage in highly NK-active F1 hosts. These 'NK-selected' sublines showed a permanently decreased sensitivity to NK lysis after 8-10 selections. When compared for in vivo growth with the parental YAC-1 tissue culture line in a spectrum of relatively resistant F1 hybrids, they had an increased frequency of takes. This is in line with recent findings which show a relationship between the target site for natural antibodies and anti-MCSA on the one hand, and between the natural antibody-binding site and the NK target site on the other.

Role of natural killer cells in control of cancer metastasis

There is a close association between levels of natural killer (NK) cell activity and the ability of the host to eliminate circulating tumor cell emboli. Mice that exhibit low levels of NK-cell-mediated cytotoxicity (3-week-old syngeneic mice, 3-week-old allogeneic nude mice, cyclophosphamide- or beta-estradiol-treated mice, and beige mice) also exhibit enhanced survival of tumor cells in the vascular bed of the lung and increased incidence of pulmonary tumor metastasis. Conversely, hosts with high NK cell activity (adult nude mice and syngeneic mice treated with NK-cell-stimulating biological response modifiers (BRM) ) are very resistant to metastasis. Lymphoid adoptive transfer studies have shown that the effector cell responsible for the antimetastatic activity is the NK cell. In these studies, NK cells were highly effective in destroying circulating tumor cells before their extravasation into the organ parenchyma, whereas they exerted only a minimal inhibiting effect on already established micrometastases. The ability to activate NK cells selectively (without subsequently inducing suppressor macrophages) provides a valuable tool for the evaluation of the role of activated NK cells in therapy of tumor metastasis. The validity of this approach is supported by the finding that NK cells activated by BRM are effective in killing, both in vivo and in vitro, solid tumor cells that developed NK-cell-resistance as a result of adaptive growth in vivo or selection during the metastatic process. An understanding of the mechanisms that regulate NK cell activation or suppression as well as elucidation of the circulatory patterns and anatomical compartmentalization of activated NK cells will help achieve a sustained systemic and/or in situ activation of NK cells which may prove effective in the control of cancer metastasis.

Incidence and type of tumors induced in C57BL bg/bg mice and +/bg littermates by oral administration of DMBA

As an attempt to study the effect of the beige (bg) mutation on chemical carcinogenesis, 65 C57Bl/bg/bg mice and 83 +/bg littermate controls received DMBA in five weekly intragastric doses. The incidence of tumors of different histological types was monitored through observation periods ranging between 165 and 500 days. By 165 days after the first DMBA feeding, 18% of the +/bg and 31% of the bg/bg mice had developed tumors. The beige mice had a higher incidence of epithelial and non-epithelial tumors arising in cutaneous or subcutaneous sites than the controls. The total incidence of lymphomas was similar in the two groups. However, lymphomas appeared somewhat earlier in beige than in control mice. Altogether 33 +/bg and 27 bg/bg mice were followed for 500 days. By this time, 73% of the +/bg and 78% of the bg/bg mice had developed tumors. The beige group showed a higher incidence of non-thymic lymphomas than the controls. In contrast, the incidence of thymic lymphoma, cutaneous epithelial tumors and bile-duct adenomas was similar in the two groups or higher in +/bg that in bg/bg mice. The results suggest that the bg mutation causes a certain defect in a mechanism that may prevent or delay the onset of non-thymic lymphomas and of epithelial and non-epithelial cutaneous tumors in DMBA-treated mice. The differences between the two groups were smaller than those previously reported in relation to the increased susceptibility of beige mice to certain transplanted tumors, attributed to the known defect in natural killer (NK) activity in the beige mice. The reduced differential in the DMBA system may be due to the partial reduction of NK activity, induced by the carcinogen, as reported previously (Ehrlich et al., 1980) and confirmed in the present study.

Quantitative studies of natural immunity to solid tumours in rats. NK activity in animals with primary or transplanted spontaneous tumours

Natural killer (NK) cell activity was measured in a quantitative 6 h chromium release assay using sarcoma MC7 cells as targets. The total NK lytic activity present in the spleen and blood of tumour-bearing animals was compared with the corresponding values for age/sex/parity-matched animals. Rats with primary spontaneous tumours in the breast or subcutaneous sites showed normal levels of NK activity, while rats with primary spontaneous kidney tumours had elevated NK activity, the degree of augmentation being greater with increasing tumour size. A similar elevation of NK activity was generally found in animals with large, transplanted, spontaneous or chemically-induced tumours. This augmentation could only detected when total lytic activity was considered: when NK activity was measured merely on a cell-for-cell basis, it often appeared to be depressed in such animals, in agreement with previous reports. However, with one rapidly metastasizing spontaneous tumour, a real depression of both spleen and blood NK activity was found. Small inocula of cells from non-immunogenic spontaneous mammary tumours or from other non-immunogenic spontaneous tumours caused no early increase in systemic NK activity when injected into the mammary pad, a site where spontaneous tumours frequently arise. However, cells from one immunogenic spontaneous tumour and 2/3 immunogenic chemically-induced tumours did occasionally stimulate significant early increases in NK activity when placed at this site. Early changes in peritoneal exudate NK activity were also investigated using small inocula of these tumours injected intraperitoneally. Augmentation of NK activity occurred with a 3-fold greater frequency following inoculation with immunogenic tumour cells than with non-immunogenic cells in this system. It can be concluded from these studies that: (1) spontaneous tumours do not selectively arise in members of an inbred strain with subnormal NK activity; (2) most large tumours in rats stimulate rather than depress NK activity; (3) early boosting of NK activity by small inocula of tumour cells placed in the mammary pad does not occur with non-immunogenic spontaneous tumours; (4) early boosting of NK activity in the peritoneal site does occur with non-immunogenic tumours, but with a very low frequency. The latter findings suggest that developing spontaneous tumours are unlikely to stimulate the NK system, and emphasize the importance of using syngeneic, spontaneous tumours for studying tumour-host relationships in animals.

Natural killer cells: their roles in defenses against disease

Natural killer cells are a recently discovered subpopulation of lymphoid cells that are present in most normal individuals of a range of mammalian and avian species. Natural killer cells have spontaneous cytolytic activity against a variety of tumor cells and some normal cells, and their reactivity can be rapidly augmented by interferon. They have characteristics distinct from other types of lymphoid cells and are closely associated with large granular lymphocytes, which comprise about 5 percent of blood or splenic leukocytes. There is increasing evidence that natural killer cells, with the ability to mediate natural resistance against tumors in vivo, certain virus and other microbial diseases, and bone marrow transplants, may play an important role in immune surveillance.

Tumour cells in metastatic deposits with altered sensitivity to natural killer cells

If natural killer (NK) cells play a role in immunosurveillance it might by expected that, during the metastatic process, selection would occur for tumour cells with reduced NK sensitivity. This hypothesis was tested in the rat by measuring the NK sensitivity of cells freshly isolated from metastases of syngeneic transplanted spontaneous mammary carcinomas. Lysis was measured in a 6-h chromium release assay using normal syngeneic spleen cells as effectors. Our studies led to the following conclusions. (1) Metastases developing at certain tissue sites (draining lymph node and lung, but not pericardium) were frequently composed of tumour cells with markedly reduced sensitivity to NK cells. (2) This resistance could generally be detected only if freshly isolated tumour cell population were studied; after a few days in culture, resistant metastasis-derived tumour cells usually regained normal NK sensitivity. (3) Resistance to NK cells was not always due to the loss of NK target structures; it could also result from an innate resistance to the NK lytic mechanism. (4) The tissue distribution of NK-resistant metastases suggested that if NK cells exerted an immunoselective pressure they did so at the tissue site rather than in the primary tumour or in the bloodstream.

Natural killing and antibody-dependent cellular cytotoxicity in specific-pathogen-free miniature swine and germ-free piglets. II. Ontogenic development development of NK and ADCC

The ontogenic development of natural killing (NK) and antibody-dependent cellular cytotoxicity (ADCC) in germ-free and specific-pathogen-fee (SPF) miniature swine were compared. Activities of NK and ADCC were tested by a short-term (2.5 to 4 h) 51Cr-labelled human myeloid cell line K562 and TNP-conjugated human B-cell line SB as target cells for NK and ADCC, respectively. Animals obtained by aseptic hysterectomy 3-5 days prior to term showed ADCC activities similar to adult levels but lacked NK activity. Hysterectomy-derived piglets which were colostrum-deprived and maintained in germ-free isolators developed NK activity at 3-4 weeks of age. In comparison, naturally-farrowed, colostrum-fed piglets maintained in our SPF facility developed NK activity at 2-3 weeks of age. Thereafter, there was no significant difference in the levels of either NK or ADCC between germ-free and SPF animals. This suggests that microbial flora and environment do not affect the development of effector cells for ADCC but do play some role in the maturation of NK cells during ontogeny. The difference in ontogeny of NK an ADCC further support our previous suggestion that the effector cells for NK and ADCC in swine are distinct sub-populations.

Increased susceptibility to cytomegalovirus infection in beige mutant mice

Mice homozygous for the beige gene (bg/bg) are a homologue of the Chédiak-Higashi syndrome of man and are known to be selectively defective in natural killer (NK) cells. We have compared the susceptibility of bg/bg and bg/+ C57BL/6J mice to infection with murine cytomegalovirus (MCMV). Beige mice are more susceptible to lethal infection and develop 33- to 43-fold higher virus titers in the liver, spleen, and kidney than do bg/+ mice after a sublethal infection, although virus replication is the same in vitro in cultured fibroblasts or epithelial cells from these mice. Inoculation with a sublethal dose of virus stimulates a NK cell response, although this is lower in bg/bg mice despite higher titers of interferon type 1 than in bg/+. A dose of MCMV that is lethal only to bg/bg augments cytotoxicity within 12 hr in bg/+ mice, whereas cytotoxicity in bg/bg remains very low. In bone marrow chimeras, recipients of bg/bg marrow were more susceptible to MCMV and had lower NK cell responses after virus inoculation than did recipients of marrow from bg/+ donors. The greater susceptibility of beige mice to the virus suggests that NK cells may contribute to resistance early in McMV infection.

Mice treated with strontium 90: an animal model deficient in NK cells

Treatment of BALB/c mice with radioactive isotopes of the bone-seeking element strontium reduces the percentage of specific NK-cell cytotoxicity to only 2.6%, compared with 13.6% for normal BALB/c and 36.3% for athymic (nude) BALB/c. The syngeneic plasmacytoma NS-1 was used as target in a 4th in vitro NK-cell microassay. Marrow cellularity in treated mice is reduced to 12.5% of controls, but haemopoietic and stem-cell functions are taken over by the spleen and the peripheral blood picture remains relatively normal. Allogenic (H-2k) tumour transplants are rejected normally with good anti-H-2k alloantibody response. Haemopoietic and T- and B-cell functions are therefore substantially intact, and the defect seems confined to NK cells. In vivo, after s.c. inoculation of 10(6) NS-1 cells, 8/12 controls grew a solid tumour after a mean delay of 30.5 +/- 1.25 (s.e.) days, whereas 5/6 90Sr-treated mice grew the tumours after a delay of only 10.5 +/- 1.8 days. This markedly reduced delay in the 90Sr-treated mice lends support to suggestions that NK cells play an important role in resisting the establishment of tumour foci (i.e. in antitumour surveillance). Mice treated with 90Sr could be useful in evaluating the in vitro role of NK cells.