Generation of lymphokine-activated killer cell activity from non-NK precursor cells

It is possible to generate high levels of lymphokine-activated killer (LAK) activity in short-term culture from cells enriched for natural killer (NK) activity. To determine whether LAK activity can also be generated from non-NK cells, we have depleted peripheral blood lymphocytes (PBL) of NK cells prior to culture with IL-2. NK activity in PBL is correlated with the intensity of staining with the lysosomotropic vital dye quinacrine. Quinacrine dim PBL, which are devoid of lytic NK cells, are capable of developing LAK activity following culture with IL-2. We have also separated PBL using the NK-associated NKH-1 marker. Depleting NKH-1+ cells eliminates NK activity but the ability to develop LAK activity is retained. NKH-1-depleted cells generate less LAK activity than unseparated or NKH-1-positive cells and do not proliferate as well as unseparated cells to IL-2. When NK-depleted cells are subsequently examined for the expression of the NKH-1 antigen, this marker is absent from most cells at Day 3 of IL-1 culture, but is expressed on an increasing number of cells by Days 6-8. These results suggest that LAK derived from non-NK cells is functionally and phenotypically similar to LAK from PBL-containing NK cells, and may be the result of the activation of an NK precursor population.     

Induction of lymphokine-activated killer cells from rat thymocytes using recombinant human interleukin-2

Summary Using a 4-h ⁵¹Cr release assay, we observed that thymocytes from Fischer strain rats incubated with recombinant human interleukin-2 (rhIL-2) developed cytotoxicity to YAC-1 lymphoma, 9L-glioma, and B-16 melanoma cells (effector/target ratio =25/1). Induction of the lymphokine-activated killer (LAK) cells was as follows: (1) when 5×10⁶/ml thymocytes were cultured with various concentrations of rhIL-2 (50, 125, 250, 500, or 1,000 units/ml) for 4 days, no cell proliferation was observed at any concentration. However, the LAK cells showed significant cytotoxicity toward all tumor cells at more than 50 units/ml. (2) When 5×10⁶/ml thymocytes were cultured for 1 to 6 days with 250 units/ml of rhIL-2, the harvested cell count decreased markedly after the 2nd day. The cytotoxicity of all the tumor cells became significant after the 2nd day, with peak activity on the 4th day. In rat splenocytes, on the other hand, the LAK cells could not be identified because rat splenocytes developed nonspecific cytotoxicity in medium containing fetal calf serum without adding rhIL-2.     

Generation of lymphokine-activated killer cell activity by low-dose recombinant interleukin-2 and tumor cells

The generation of lymphokine-activated killer (LAK) cells in vitro has been reported to require 100-1000 units of recombinant interleukin-2 (IL2). In this study we investigated the generation of human LAK cells with low-dose IL2 (1-10 U) in combination with human tumor cell lines. A significant LAK activity was generated within 3- to 5-days culture of PBL. Among six human tumor cell lines tested, the K562 cell line had the greatest stimulating activity, and the degree of cytotoxicity was comparative to that of PBL stimulated with higher doses of IL2 alone. The origin of this LAK activity was primarily the E(-) rosetting cell population. Cocultures of E- cells with 1 U/ml IL2 plus K562 had significantly higher cytotoxicity (P less than 0.05) compared to using E+ cells. Phenotypic analysis indicated that 1 U/ml IL2 plus K562 cell stimulation enhanced CD56+ and CD16+ cells. These studies suggest that very low dosages of IL2 with stimulator tumor cells can generate LAK activity comparable to that generated with high dosages of IL2 alone.     

Cytostatic and cytotoxic activity of lymphokine-activated killer cell supernatants

Summary Incubation of peripheral blood mononuclear cells with interleukin-2 (IL-2) results in the release of a factor which is cytostatic and cytotoxic both to tumor cell lines (A375M, A375P, C480, MCF-7, Hey) and fresh tumor cells (in the human tumor cloning assay), including breast cancer, colon cancer, melanoma, myeloma and ovarian cancer. The factor cannot be detected in a 4-h chromium-release assay, but is best demonstrated after tumor cells have been to it for exposed 3 days. The factor is not cytotoxic to normal peripheral blood leukocytes or normal fibroblasts, and is not toxic to certain targets sensitive to lymphokine-activated killer (LAK) cells, such as K562 and Daudi cells. The factor is diffusible, non-dialyzable, relatively stable to heat and acid and does not contain appreciable amounts of targets resistant to interferon- and , tumor necrosis factor and interleukin-1. The data suggest that there are several mechanisms of LAK cell activity against tumor cells including one which requires direct interaction of LAK and tumor cells and one which is mediated by LAK cell supernatant. The former is detected by 4-h chromium release while the latter is not.Supported by a grant from the Arizona Disease Research Commission and by Grants CA-17094 and CA-23074 from the National Institute of Health     

IL-4 regulates IL-2 induction of lymphokine-activated killer activity from human lymphocytes

IL-4 is a pluripotent lymphokine acting on various cell types. We investigated the role of human IL-4 on the generation of lymphokine-activated killer (LAK) activity. Human IL-4 alone did not induce LAK activity and inhibited IL-2 induction of LAK activity from unstimulated PBMC, peripheral blood null cells, spleen cells, and lymph node cells in a dose-dependent manner. IL-4 also inhibited several phenomena induced by IL-2 such as cell proliferation, augmentation of NK activity, increase of Leu-19+ cells, and expression of IL-2R(p55) on either CD3+ or Leu-19+ cells. IL-4, however, augmented cell proliferation with other T cell mitogens including PHA, Con A, PMA, or allo-MHC Ag with or without IL-2. In contrast to unstimulated cells, IL-4 alone induced marked cell proliferation and LAK activity as well as Leu-19+ cells from in vitro IL-2 preactivated PBMC or null cells, and did not inhibit IL-2 induced cell proliferation, LAK activity, Leu-19+ cells and IL-2R(p55) expression, but rather augmented them with low doses of IL-2. Although IL-4 alone induced LAK activity from peripheral blood of some patients previously given IL-2, IL-4 inhibited in vitro LAK generation with IL-2 from these cells in most cases. Therefore, IL-4 appears to directly inhibit the IL-2 activation pathway via IL-2R(p70) and prevent resting LAK precursors from proliferating and differentiating into final effector cells. However, once cells were sufficiently preactivated by IL-2, IL-4 induced LAK activity and did not inhibit IL-2 activation of these cells. These data suggest an immunoregulatory role of IL-4 on human null cells and T cells.     

Generation of lymphokine-activated killer (LAK) cells from tumor-infiltrating lymphocytes

Culture of tumor-infiltrating lymphocytes (TIL) containing about 20% BMC2 tumor cells with recombinant human interleukin 2 (rIL-2) resulted in the diminish of tumor cells and the growth of lymphocytes. These IL-2-activated lymphocytes showed a strong cytotoxic activity against not only syngeneic tumor cells but also allogeneic tumor cells. Such broad-reactive killer cells, termed lymphokine-activated killer (LAK) cells, are also inducible from spleen cells by in vitro activation with IL-2. However, LAK cells generated from TIL (TIL-LAK) showed higher cytotoxic activity against BMC2 than LAK cells generated from spleen cells (S-LAK). Furthermore, it was demonstrated that TIL-LAK cells revealed marginal cytotoxic activity against normal Con A blasts and YAC-1 cells as opposed to S-LAK. Flow cytometric analysis of TIL-LAK indicated that TIL-LAK cells mainly consisted of Thy 1.2+, Ly 2+, asialo GM1+ cells. TIL-LAK cells displayed not only in vitro cytotoxicity but also in vivo anti-tumor activity. Furthermore, it was also confirmed that TIL-LAK cells could be induced in autochthonous mouse tumor systems and human gastric tumor systems.     

Generation of lymphokine-activated killer activity in T cells. Possible regulatory circuits.

CD4+ and CD8+ T cells do not develop significant lymphokine-activated killer (LAK) activity when PBL are cultured with IL-2 or even when they are activated with a T cell stimulus such as OKT3 mAb. The possibility that a T cell regulatory mechanism prevents the development of LAK activity by CD4+ or CD8+ cells in OKT3 mAb and IL-2 cultures was tested by depleting CD8+ or CD4+ cells from PBL before stimulation with OKT3 and IL-2. Under these conditions, the remaining CD4+ and CD8+ cells were able to generate non-MHC-restricted lysis of NK-resistant tumor targets. Our data suggested that a regulatory signal was present in the culture to prevent the development of lytic function by T cells. T cells removed from the PBL cultures were, upon culture with IL-2, able to generate high LAK activity, suggesting that inhibition of the CD4+ or CD8+ T cell-mediated LAK activity was an active ongoing process, which blocked the lysis at the level of the activated cell and not the precursor cell. Mixing experiments demonstrated that the CD4+ or the CD8+ cells isolated from the PBL cultures were able to inhibit the development of lytic function in the CD4-depleted and CD8-depleted cultures. Transforming growth factor-beta (TGF-beta) has been shown to block LAK activity of NK cells in IL-2-stimulated cultures. When TGF-beta was added to CD4(+)- or CD8(+)-depleted cultures, it also inhibited LAK activity of T cells in a dose-dependent fashion, without interfering with T cell growth. Lytic activity returned to activated levels when TGF-beta was removed from the culture medium, thereby demonstrating the reversibility of TGF-beta inhibition.     

IL-7 induces human lymphokine-activated killer cell activity and is regulated by IL-4.

Induction of lymphokine-activated killer (LAK) activity by IL-2 has been described and characterized as broadly cytolytic activity against both fresh and cultured tumors. rIL-7 in the absence of IL-2 also induces LAK activity in human cells. This activity is unique for IL-7, because it is not shared by other cytokines including IL-1, IL-4, IL-6, and TNF-alpha. IL-7 also induces either de novo or increased expression of the surface markers CD25 (Tac, IL-2R alpha-chain), CD54 (ICAM-1), Mic beta 1 (IL-2R beta-chain) and CD69 (early T cell activation Ag). IL-7-induced LAK activity is independent of IL-2 secretion, because it is not abrogated by IL-2 antisera. The LAK precursor responding to IL-7 stimulation is enriched in the null cell fraction as has been demonstrated for IL-2-induced LAK cells. TGF-beta and IL-4 interfere with generation of LAK activity by IL-7. Anti-IL-4 antiserum enhances IL-7-induced LAK activity and augments induction of surface marker expression by IL-7. This may be indirect evidence that IL-7 stimulation leads to induction of IL-4 activity. Our results describe the activation of mature lymphoid cells by IL-7. This and the previously described role of IL-7 in lymphohemopoiesis makes it a cytokine of potential therapeutic value for treatment of immunodeficiency states and possibly the immunotherapy of cancer.     

Heterogeneous lymphokine-activated killer cell precursor populations

Summary We developed a monoclonal antibody (mAb) 211, which recognizes the precursors in peripheral blood of lymphokine-activated killer cells (LAK) induced by recombinant interleukin-2 (rIL-2). In conjunction with complement mAb 211 also eliminates natural killer cells (NK) and a majority of the cytotoxic T lymphocytes. B cells and monocytes do not express the 211 antigen. Since mAb 211 recognized such a large percentage of peripheral blood lymphocytes we examined which 211⁺ subpopulation was the predominant precursor of rIL-2-induced LAK cells using two-color fluoresence-activated cell sorting (fluorescein-conjugated 211 mAb plus phycoerythrin-CD11b). This method identified the 211⁺/ CD11b⁺ population as the predominant phenotype of the rIL-2-induced LAK precursor. In addition, we directly compared the phenotype of the LAK precursor induced by delectinated T-cell growth factor (TCGF) to that induced by rIL-2. The 211-depleted population, which was devoid of NK cells and LAK precursors (inducible by rIL-2), was capable of generating LAK activity when TCGF was used as the source of lymphokine. LAK cells induced by TCGF from the 211-depleted population lysed a fresh sarcoma and an NK-resistant cultured melanoma tumor target but not the Daudi cell line, which was lysed by rIL-2-induced LAK cells. Lymphoid subpopulations, depleted using NKH1a mAb, behaved similarly, generating high levels of lysis against the two solid tumor targets when cultured with TCGF but not with rIL-2. CD 3-depleted populations showed enrichment for LAK precursors using either rIL-2 or TCGF. These results indicate that while rIL-2-induced LAK precursors cannot be separated from cells with NK activity, TCGF-induced LAK cells can be generated from populations of peripheral blood mononuclear cells without NK activity.     

Induction of lymphokine-activated killer cell against human leukemia cells in vitro

The induction of lymphokine-activated killer (LAK) cells against fresh human leukemia cells was investigated. Two thirds of the 62 leukemias examined were susceptible to the lytic effect of allogeneic IL-2 induced LAK cells in vitro. No substantial differences could be detected between myeloid or lymphoid leukemias or with regard to the FAB subtype or the immunophenotype. Culturing mononuclear cells from peripheral blood or bone marrow of leukemia patients with IL-2 resulted in an expansion of residual large granular lymphocytes and development of cytotoxic activity. The combination of IL-2 with IFN-gamma or the presence of tumor cells during the activation process led to an enhancement of LAK cell cytotoxicity. These results suggest that LAK cells may be useful in the treatment of leukemia.     

Chemotherapy-induced modulation of natural killer and lymphokine-activated killer cell activity in euthymic and athymic mice

Combinations of chemotherapy and interleukin-2 (IL-2) aimed at improving therapeutic efficacy in cancer patients have generally proved disappointing. Although chemotherapy blocks tumor growth and sometimes boosts immune functions, most drugs are immunosuppressive, at least transiently. Therefore, it is reasonable to assume that maximal exploitation of the immunostimulatory and antitumor activity of both modalities requires careful coordination of chemotherapy and IL-2 timing. We analyzed the temporal effect of 5-fluorouracil (5-FU, 100–120 mg/kg), cyclophosphamide (CY, 100 mg/kg), Adriamycin (8 mg/kg) and dacarbazine (100 mg/kg) on the activation of natural killer/lymphokine-activated killer (NK/LAK) cells by IL-2 in several strains of euthymic mice and in athymic nude mice. Following in vivo or in vitro exposure to IL-2 1–15 days after chemotherapy, the total lytic activity of the spleen and the number of LAK precursors (LAK-p) were measured. In euthymic mice injected with IL-2 (5×10⁴ Cetus units twice daily for 4–5 days), 5-FU augmented (up to 37-fold, days 1–9) and CY reduced (up to day 6) LAK activity, as compared with that in the IL-2 control. In bulk cultures containing IL-2 (1000 CU/ml, 3–4 days), both 5-FU and CY reduced LAK activity of euthymic mice splenocytes for up to 6 days after chemotherapy, which was followed on day 9 by full recovery. In splenocytes of nude mice, 5-FU increased and CY diminished LAK activation in bulk cultures, starting 3 days after chemotherapy. In athymic mice, 5-FU markedly augmented the total number of LAK-p/spleen (up to 30-fold, days 3–9), as determined by limiting-dilution cultures with IL-2 (for 7–8 days). In euthymic mice, in contrast, LAK-p levels decreased for up to 6–9 days after treatment with 5-FU, Adriamycin or dacarbazine, later recovering to pretreatment levels, whereas CY markedly increased LAK-p (up to 15-fold) when administered 6–12 days before limiting-dilution culture initiation. The effect of chemotherapy on LAK and NK activity was essentially similar. In other experiments, a subset of asialoGM1-LAK-p was found in the spleens of 5-FU-treated mice, but not in untreated mice. Our results suggest that the immunomodulatory effect of chemotherapy on NK/LAK activity in mice is variable and largely depends on the drug itself, the interval between chemotherapy and IL-2 administration, the strain of mice and the assay used.     

Lysosome rich cells contain the lytic activity of lymphokine-activated killer cell populations

Summary Little is known regarding the effectors of lymphokine-activated killer activity. Lysosomotropic agents such as quinacrine can be used to positively sort for lysosome rich cells in natural killer (NK) cell populations. We therefore decided to use this agent to sort lymphokine-activated killer (LAK) cells to characterize their lysosomal content. We found that the positively sorted population contained all the LAK activity, i.e., lysis of NK-resistant tumor cells (B16 melanoma cell line), with the negatively sorted cells having no killing activity. Therefore separation of interleukin-2-incubated cells for LAK activity could be accomplished using sorting after quinacrine staining. The treatment of positively sorted LAK cell populations with L-leucine methyl ester, a lysosomotropic dye which inhibits killing by lysosome rich cells, caused abrogation of killing of the B16 tumor by the treated populations. Single cell conjugate assays were also done on these sorted cells, with positively sorted cells forming the highest and negatively sorted cells the lowest percent of conjugates. Our data therefore indicates the important role of lysosome rich cells in the LAK cell population in the murine system.This work was supported by NIH grants R01 CA42962 and K04 CA0122, and by intramural funds from the Norris Cancer Center     

Generation of long-term human cytolytic cell lines with persistent natural killer activity

Human cell lines maintained by in vitro stimulation with the HLA-A, B-negative, DR-positive, Epstein Barr virus-transformed, lymphoblastoid cell line Daudi in the presence of conditioned medium demonstrated significant NK activity for over 6 wk in continuous culture. These cells lyse K562 and a broad panel of lymphoblastoid cell lines but do not lyse normal peripheral blood lymphocytes or pokeweed mitogen blasts. They possess the sheep red blood cell receptor but lack other T cell markers (Lyt-3+, OKT3-). Natural killer activity correlated with the presence of a Mac 1-positive subpopulation of cells present in these long-term lines.     

The effects of staphylococcal protein A on human lymphokine-activated killer cell induction

Summary A murine IgG2b monoclonal antibody directed to the constant part of the human / T cell receptor (BMA031) was investigated in a pilot study as an initial treatment for acute graft-versus-host disease (aGvHD) after allogeneic bone marrow transplantation. The treatment protocol consisted of 5 mg BMA031 on 5 consecutive days with continuation of the prophylactic baseline immuno suppression using cyclosporin. Seven patients with grades II–III acute graft-versus-host disease were entered on the protocol and six patients completed the full treatment course. Mild to moderate acute adverse reactions to the first BMA031 infusion occurred in three patients. A nearly complete decline of circulating T lymphocytes was observed during BMA031 therapy, but the T cells returned to pretreatment values within 1 week after the last infusion. Serum pharmacokinetics of free antibody best fitted to a two-compartment open model with a mean initial half-life of 6 h and an estimated mean terminal half-life of 40 h. One patient developed antimurine antibodies of the IgM subclass. In five patients a complete and sustained resolution of all disease manifestations was attained, while in one patient a temporary response of skin involvement with aGvHD was noted. These results indicate that BMA031 can be safely administered as initial treatment of aGvHD. The therapeutic responses observed warrant its further clinical evaluation in this setting.     

Phenotypic characterization of lymphokine-activated killer cells from human lymph node lymphocytes

We previously reported that lymphokine-activated killer (LAK) activity can be generated in human lymph node lymphocytes (LNL) at the same level as that in peripheral blood lymphocytes (PBL), despite the absence of active natural killer (NK) cells. In the present study, we investigated the surface phenotype of LNL-LAK cells by fractionation of lymphocytes, using a panning method. LNL isolated from lung cancer patients were cultured in the presence of recombinant interleukin 2 for 8 days and separated into T cells and non-T cells according to the expression of CD3 antigen. LAK effectors were enriched in the CD3- non-T cells. However, the CD3+ cells also mediated a low but substantial level of LAK activity, which was attributed to a CD8+ T-cell subset. Further investigation of the CD3- cells revealed that most of the CD3- effector cells expressed neither B-cell (CD20) nor NK-cell (CD16) markers. Precursors of this CD3-CD20-CD16- (null) population appeared to be also CD3-, CD20-, and CD16-. From these results, we would stress the significant contribution of CD3-CD20-CD16- null cells to the LAK phenomenon, which has not been focused on in PBL.     

Regulation of lymphokine-activated killer cell induction by human recombinant IL-1 receptor antagonist. Obligate paracrine pathway of IL-1 during lymphokine-activated killer cell induction.

IL-2-stimulated human lymphocytes, referred to as lymphokine-activated killer (LAK) cells, can develop a broad range of lytic activity against fresh tumor cells and cultured tumor cell lines. IL-1, a pleiotropic cytokine shown to synergize with IL-2 on LAK induction, is endogenously synthesized and secreted by LAK cells. Immunoblot analysis demonstrated that IL-2-stimulated PBL produced the 31- to 34-kDa pro-molecules of IL-1 within 24 h and maintained their expression for at least 96 h. The role of secreted IL-1 has been examined using rIL-1R antagonist (IL-1ra). The addition of IL-1ra to LAK activation culture resulted in dose-dependent inhibited lytic activity, which was more apparent in LAK cells cultured with higher doses of IL-2. However, IL-1ra had no effect on proliferative responses elicited in LAK cells by IL-2. Moreover, when IL-1 binding was blocked by IL-1ra, the expression of the IL-2R p55 subunit was reduced compared with control LAK cells. The effect of IL-1 binding blockade on expression of other cytokine mRNA was further examined by polymerase chain reaction analysis, and, specifically, inhibition of both TNF-alpha and TNF-beta mRNA expression by IL-1ra was observed in PBL stimulated with IL-2. The reduced biologic activity of TNF in culture supernatants correlated well with the inhibition of mRNA expression. These findings suggest that autocrine/paracrine IL-1 is involved in the initial generation of LAK activity and, in particular, that TNF expression could be induced via an IL-1 autocrine pathway.     

Inhibition of murine lymphokine-activated killer (LAK) cell activity by adherent cells

The effects of adherent cell depletion, indomethacin, and prostaglandin E2 (PGE2) on murine LAK cell activity were investigated. Removal of plastic adherent cells from splenocyte suspensions either prior to 5-day culture with 1000 U/ml of recombinant human IL-2 (rIL-2) or prior to assay resulted in an enhanced LAK cell cytotoxicity compared to that of whole spleen cell suspensions. Indomethacin enhanced LAK cell cytotoxicity of whole splenocyte suspensions if present during the culture period, but had no effect on whole splenocyte or adherent cell-depleted cell suspensions if added just prior to assay. PGE2 suppressed LAK cell activity of nonadherent splenocyte but not whole splenocyte suspensions when present during the culture period. In vivo treatment of mice with indomethacin enhanced cytotoxicity directed toward both LAK sensitive, natural killer (NK) resistant (P-815) and LAK, NK sensitive (YAC-1) tumor cell targets. Splenocytes from indomethacin-treated mice cultured with additional indomethacin and rIL-2 exhibited highest LAK cell activity. The results from this study indicate that LAK cells are regulated by adherent cells which suppress LAK cell activity. This suppression can be reversed both in vitro and in vivo by indomethacin. This study has important implications for the possible clinical use of indomethacin in the potentiation of in vivo and in vitro LAK cell activity for immunotherapeutic protocols.     

Heterogeneity of the lymphokine-activated killer cell phenotype.

Lymphokine-activated killer cells (LAK) are functionally defined by their ability to mediate the MHC-unrestricted lysis of a range of tumor targets, while sparing normal cells. They can also lyse TNP-modified normal syngeneic lymphoblasts. We show here that lysis of TNP-modified targets is mediated by CD8+ LAK in a self-MHC-restricted manner, whereas lysis of tumor targets is largely by CD8- LAK and is MHC-unrestricted. LAK generated from the immune-deficient strains Balb/c nude and C.B-17 scid lyse tumor targets as effectively as LAK from normal mice but do not lyse TNP-modified normal targets. Further, lysis of TNP-modified targets, but not tumors, can be inhibited by antibody to the T cell receptor complex. These experiments strongly suggest that recognition of TNP-modified targets is not accomplished by the same mechanism as that of tumors. Rather, they are consistent with recognition of TNP-modified targets by CD8+ LAK cells being mediated via recognition through the T cell receptor.     

Lysis of human monocytes by lymphokine-activated killer cells

Human peripheral blood leukocytes (PBL), stimulated in vitro with recombinant human interleukin 2 (IL-2) for 2-7 days, were seen to lyse autologous and allogeneic monocytes in a 4-hr 51Cr-release assay. The lymphokine-activated killer (LAK) cells against monocytic cells were selective in that polymorphonuclear leukocytes (PMN) and nonadherent PBLs were not lysed by these cells. Monocytes which had been cultured for 2-7 days served as better targets than uncultured cells. Also, kinetic studies demonstrated parallel activation of cytolytic activity against monocyte targets and FMEX, an natural killer cell-insensitive human melanoma target. Separation of PBLs by discontinuous density centrifugation identified the effector population in the fractions enriched for large granular lymphocytes (LGL). Precursor cells were seen to express CD2, CD11, and some CD16 markers, but not CD3, CD4, CD8, CD15, Leu M3, or Leu 7. The effector population after IL-2 activation retained the phenotype of the precursor cell. These studies indicate that IL-2 can generate LAK cells against monocytic cells, and this cytolytic activity, especially against autologous monocytes, must be taken into account when IL-2 or LAK cells are used for immunomodulation in cancer patients.