Respiratory Tract Gene Transfer TRANSPLANTATION OF GENETICALLY MODIFIED T-LYMPHOCYTES DIRECTLY TO THE RESPIRATORY EPITHELIAL SURFACE*

To evaluate the strategy for potentially treating res- piratory disorders with genetically modified T-lym-phocytes, the interleukin-2 (IL-a)-dependent murine T-cell line, CTLL2, was genetically altered with the Escherichia coli 8-galactosidase (@-gal) gene (lac2) in vitro with a retroviral vector and the modified T-cells were transplanted directly to the respiratory epithelial surface of syngeneic C57B1/6 mice. Southern and Northern analyses confirmed that the neomycin-se- lected modified T-cells contained and expressed the lac2 gene. The fate of the modified T-cells (CTLL2/ lacZ) was followed by flow cytometry with T-cell surface marker Thyl.2 and fluorescent &gal analysis. One day after transplantation % T-cells/g f 3% of the Thyl.2+ T- cells recovered from respiratory epithelial lining fluid (ELF) were &gal+. Importantly, the modified T-cells remained in the lung for some time; at 3 days, Thyl.2+ &gal+ T-cells 63 f 12% of ELF Thyl.2+ T-cells and 59 f 6% of Thyl.2+ T-cells recovered from the whole lung. At 7 days, 33 f 8% of the Thy 1.2+ cells in and 75 f 6% of the Thyl.2+ cells in whole lung were Thyl.2+ ,&gal+. In contrast, the proportion of the Thyl.2+

To evaluate the strategy for potentially treating respiratory disorders with genetically modified T-lymphocytes, the interleukin-2 (IL-a)-dependent murine T-cell line, CTLL2, was genetically altered with the Escherichia coli 8-galactosidase (@-gal) gene (lac2) in vitro with a retroviral vector and the modified T-cells were transplanted directly to the respiratory epithelial surface of syngeneic C57B1/6 mice. Southern and Northern analyses confirmed that the neomycin-selected modified T-cells contained and expressed the lac2 gene. The fate of the modified T-cells (CTLL2/ lacZ) was followed by flow cytometry with T-cell surface marker Thyl.2 and fluorescent &gal analysis. One day after transplantation (7.5 % 10' CTLL2/lacZ T-cells/g of body weight), 95 f 3% of the Thyl.2+ Tcells recovered from respiratory epithelial lining fluid (ELF) were &gal+. Importantly, the modified T-cells remained in the lung for some time; at 3 days, Thyl.2+ &gal+ T-cells represented 63 f 12% of ELF Thyl.2+ T-cells and 59 f 6% of Thyl.2+ T-cells recovered from the whole lung. At 7 days, 33 f 8% of the Thy 1.2+ cells in ELF and 75 f 6% of the Thyl.2+ cells in whole lung were Thyl.2+ ,&gal+. In contrast, the proportion of the Thyl.2+ @-gal+ T-cells in the spleen, the major extrapulmonary lymphatic organ, never rose above 3 2 1% of the total Thyl.2+ cells. The number of Thyl.2+ @-gal+ T-cells in the lung could be modified by the systemic administration of IL-2, with whole lung Thyl.2+ &gal+ T-cells increasing 4.6-fold 3 days after transplantation, compared with non-IL-2-treated animals. These studies suggest that direct transplantation of genetically modified T-cells into the lung is feasible and represents a viable strategy for lung-specific gene transfer.
A variety of human disorders can potentially be treated by modifying the genetic composition of autologous T-lymphocytes in a fashion relevant to the specific disease and directing the modified T-cells to the site of disease (1-3). Although the process of inserting new genes into lymphocytes can be accomplished using vectors such as retroviruses (1)(2)(3)(4)(5)(6)(7)(8), one limitation of this treatment strategy is the problem of targeting the modified T-cells to the site of disease. Since the major manifestations of most disorders are usually limited to specific * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ T o whom reprint requests should be addressed: Pulmonary Branch, Bldg. 10, Rm. 6D03, National Institutes of Health, Bethesda, MD 20892. Tel.: 301-496-1597;Fax: 301-496-2363. anatomic locations, systemic administration of genetically modified T-lymphocytes requires that, unless the modified Tcells "home" to the disease site; otherwise potential therapeutic efficacy may be limited. This, together with the risk of having potentially "dangerous" genetically modified T-cells circulating throughout the body and reaching organs with no disease, argues strongly for targeting the modified T-cells directly to specific organs.
Theoretically, targeting of T-cells could be achieved by capitalizing on the inherent (or engineered) properties of Tcells to home specifically to the site of disease. This concept is the underlying premise of modifying tumor infiltrating lymphocytes and reinfusing the modified, expanded population of cells with the expectation they will localize to the neoplasm (6,8). Alternative strategies for homing are to capitalize on specific T-cell antigen receptors or specific homing ligands (9, 10).
In the present study, we have explored a different approach to the problem of targeting, that of directly transplanting the genetically modified T-cells to the organ of interest. As a model for this strategy, we have evaluated the consequences of directly transplanting genetically modified T-cells to the epithelial surface of the lung. To do so, the genome of the murine T-cell line CTLLB was modified using a retroviral vector carrying the Escherichia coli lac2 ((%galactosidase) gene (11-13). The modified T-cells transplanted to the respiratory epithelial surface of syngeneic C57B1/6 mice, and the relative numbers of modified T-cells on the respiratory epithelial surface and in the lung parenchyma quantified over time.

Transplantation of
Modified T-lymphocytes long terminal repeats of the Moloney murine leukemic virus and the neomycin resistance gene driven by the SV40 early promoter. The HamHI fragment of pMC1871 (Pharmacia LKB Biotechnology Inc.) containing the lacZ gene was altered at the ends and inserted into the XhoI site of pLXSN to create pLBgSN. This final construct consists of the 5' long terminal repeat driving the lac2 gene followed by the SV40 early promoter-controlled neomycin resistance gene (Fig.  1A). An infectious amphotropic virus-producing packaging cell line was produced by infecting PA317 cells (17) with supernatant from a pLBgSN transfected \k2 cell clone and selected with 1.0 mg/ml G418 (Geneticin, GIBCO). pLBgSN-infected PA317 cells were maintained in improved minimum essential medium (Biofluids, Inc.), 10% fetal calf serum, 2 mM glutamine, 100 units/ml penicillin, 100 pg/ml streptomycin, and 1.0 mg/ml G418. One of the monoclones from the pLBgSN-infected PA317 cells, pLBgSN17 (viral titer of 104-105 colony-forming units/ml), was used for all subsequent experiments.
To infect CTLL2 cells, 10 ml of infectious retroviral supernatant of the virus-producing pLBgSN17 clone was added to a suspension of 1 x 10" CTLL2 cells in 5 pg/ml Polybrene (Sigma), and 2000 units/ ml IL-2 in a 50-ml culture flask and cultured overnight at 37 "C. Retroviral supernatant was then replaced with nonselective IL-2 (2000 units/ml) containing medium for an additional 24 h. The infected cells were then selected with medium containing 1 mg/ml G418 and 2000 units/ml IL-2, and the numbers of T-cells were expanded for 2 weeks. The selected cells (CTLL2IlacZ) were maintained with 0.5 mg/ml G418 and 50 units/ml IL-2.
I n Vitro Evaluation of CTLLZ/lacZ Cells-Integration of the lacZ cDNA into the genome of CTLL2IlacZ cells was evaluated by Southern analysis following PuuII digestion of 10 pg of genomic DNA. After agarose gel electrophoresis and transfer to nylon membrane (Schleicher and Schuell), the 2.5-kb PuuII fragment of lacZ was detected using a '"P-labeled lac2 probe generated by the random primer method (Promega Biotec, Madison, WI) and subsequent autoradiography (18). Expression of lac2 transcripts were evaluated by Northern analysis following formaldehyde-agarose gel electrophoresis of 10 pg of total RNA, transfer to nylon membrane, hybridization to the '"P-labeled lacZ probe, and autoradiography (19).
To determine that CTLL2IlacZ cells maintained a normal proliferative response to IL-2, incorporation of ["Hlthymidine after incubation with various concentrations of recombinant human IL-2 (0.1-100 units/ml; Genzyme Corp., Boston, MA) was compared with that of uninfected CTLL2 cells evaluated under the same conditions (22).
Transplantation of Modified T-cells-Prior to transplantation, control CTLL2 and CTLL2IlacZ cells were kept in culture for a t least 1 week in nonselective media at a density of less than 5.0 X 105/ml, conditions which maintain viability of the cultured cells at 290% as determined by trypan blue exclusion. The cultured cells were collected by centrifugation a t 200 X g for 5 min a t 4 "C, washed twice with PBS, and resuspended in PBS at 1.5 X 10X/ml for intratracheal administration; 7.5 X 10" cells/g of body weight were delivered to the mice.
Following transplantation of CTLL2 or CTLL2IlacZ T-cells, mice were evaluated on days 1,3, 7, or 14. Following excision of the spleen, intact lungs were perfused in uiuo with PBS from the right atrium to remove cells from the vascular compartment. To assess transplanted cells populating the lung epithelial surface, bronchoalveolar lavage was performed using four 1-ml aliquots of PBS via a 20-gauge catheter and immediately gently aspirating each aliquot to obtain a total of 3.0 ml of fluid containing cells from the respiratory epithelial lining fluid (ELF). To evaluate the possibility that transplanted cells trafficked from ELF to the alveolar interstitium, T-cells were recovered from the whole lung. Briefly, following killing and vascular perfusion with PBS, lungs were excised, weighed, finely minced with scissors, and stirred in PBS and 10 mM EDTA (Sigma) a t 4 "C for 45 min (23). The minced lung was passed through mesh to recover cells, which were collected by centrifugation at 200 X g for 5 min and resuspended in "staining media" (4% fetal bovine serum, 10 mM HEPES, pH 7.0 (Biofluids, Inc.) in PBS) for analysis by flow cytometry. For convenience, the cells recovered from the respiratory epithelial fluid will be referred to as cells in ELF, and the cells extracted from the minced lung will he referred to as cells in whole lung. To evaluate CTTL2IlacZ cell trafficking to extrapulmonary lymphoid tissue, the spleen was removed and compressed in a tissue culture dish, and the cells were recovered as described for the whole lung.
To evaluate the location of transplanted CTLL2IlacZ cells in the lung, staining of Thyl.2 and &gal was performed on cryostat mouse lung sections. The lungs were perfused with PBS from the right atrium and then with 4% paraformaldehyde. Fixed lungs were excised, submerged in 4% paraformaldehyde overnight, and then frozen. Cryostat sections were first stained with biotin-labeled monoclonal mouse antibody to Thyl.2 followed by peroxidase-labeled avidinbiotin complex (Vector Laboratories, Burlingame, CA) and then incubated a t 23 "C for 5 min with chromogen 3,3'-diaminobetidine (Sigma), yielding a brown reaction product. Lung sections were then stained for the blue &gal product using X-gal as described above.
Quantification of Thyl.Z+ B-Gal+ T-cells-Flow cytometry analysis was used to simultaneously quantify T-cell surface phenotype Thyl.2 and cytosolic @-gal in CTLL2IlacZ T-cells recovered in respiratory ELF, whole lung, and spleen. Following cell recovery, a minimum of 1 X lo6 cells were pelleted and resuspended in 100 pl of staining media. The cells were stained for the T-cell surface marker Thyl.2 by incubating (30 min, 4 "C) with biotin-conjugated anti-mouse Thyl.2 monoclonal antibody (Becton Dickinson, Immunocytometry Systems, Mountain View, CA), washing with staining media, and incubating with phycoerythrin-conjugated streptavidin (Becton Dickinson) for 30 min a t 4 "C. As a control irrelevant antibody, CTLL2 or CTLL2IlacZ cells were incubated with a monoclonal antibody to murine helper T-cell phenotype L3T4 (Becton Dickinson). For the measurement of @gal activity, the cells were next warmed (37 "C, 10 min) and then exposed to 1 mM fluorescein di-P-galactopyranoside (FDG; Molecular Probes, Inc., Eugene, OR) for 1 min, after which the reaction was stopped by adding ice-cold staining media (13). Propidium iodide was added to stain dead cells for exclusion by flow cytometry (24). The cells were kept at 4 "C until flow cytometry was performed.
Flow cytometry analysis was carried out using FACScan (Becton Dickinson). The settings were kept constant for all samples and a t least lo4 events were counted. CTLL2 and CTLL2IlacZ cells stained with Thyl.2-phycoerythrin and FDG were used to set statistical quadrants so that >96% of the Thyl.2+ &gal-cells were in the upper left quadrant and >96% of the Thyl.2+ P-gal+ cells were in the upper right quadrant. Spleen cells from normal mice were used to adjust the forward scatter lower threshold so that all Thyl.2+ cells were analyzed. Background, nonspecific FDG staining in the samples from ELF or whole lung was determined by analyzing cells from control experiments in which noninfected CTLL2 cells (Thyl.2+ Bgal-) were transplanted via the trachea and recovered from ELF and whole lung a t 1 and 7 days. The relative number of cells falling into the two statistical regions was then compared: Thy 1.2+ B-gal-(left upper quadrant) and Thyl.2+ @-gal+ (right upper quadrant). The linear relationship ( r = 0.98) found between these two quadrants was used to subtract the random FDG background from samples of CTLL2IlacZ cells recovered in uiuo. The number of Thyl.2+ P-gal+ T-lymphocytes was determined by multiplying the number of total cells recovered by the percentage of Thyl.2+ P-gal+ cells obtained.
I n Viuo Expansion of the Number of CTLLZIlacZ Cells-To assess whether systemic administration of IL-2 modified the numbers of Thyl.2+ P-gal+ T-cells in the lung following intratracheal transplantation of the CTLL2IlacZ T-cells, 10,000 units of human recombinant IL-2 was administered via the peritoneum to mice 3 times daily (25) after intratracheal transplantation of 7.5 X lo5 CTLL2IlacZ T-cells/ (g of body weight). On day 3, cells from ELF, whole lung, and spleen of IL-2-treated and -untreated mice were analyzed by flow cytometry as described above. To confirm that the IL-2 influenced the numbers of transplanted CTLL2IlacZ cells in the lung (and not simply the ability to recover the cells), total RNA was extracted from homogenized whole lung 3 days post-transplantation. The extracted RNA (20 pg for each animal) was evaluated by slot blot analysis, hybridization with a "'P-labeled lac2 probe, and autoradiography (19).
Statistical Eualuation-All data are expressed as mean k standard error of the mean. All statistical analyses were carried out using the two-tailed Student's t test.

In Vitro Analysis of T-cells Genetically Modified with la&-
Southern, Northern, and protein analyses confhned that the CTLL2IlacZ T-cells contained an integrated, functional lac2 gene (Fig. 1). In this regard, genomic DNA of CTLL2 cells evaluated by Southern analysis with a 32P-labeled lac2 probe demonstrated no lac2 sequence, whereas a 2.5-kb PuuII lac2 fragment was clearly evident in the genome of CTLL2IlacZ T-cells (Panel B ) . Northern analysis using a 32P-labeled lac2 probe demonstrated no lac2 mRNA transcripts in CTLL2 cells but that CTLL2IlacZ cells contained the expected 6. CTLL2 cells using a 0-gal colorimetric assay was negative, but &gal activity was clearly present in CTLL2IlacZ cells; 0gal activity in CTLL2IlacZ was 14.3 x 10" units/106 cells, corresponding to 2.4 X lo5 molecules of P-gal/cell(20). Finally, the lac2 protein product, /%gal, was not detected by cytochemical staining with X-gal in the CTLL2 T-cells, but was readily detected in the CTLL2IlacZ T-cells (Panels D and E ) . Flow cytometric analysis comparison of CTLL2 and CTLL2IlacZ cells demonstrated that almost all of the CTLL2IlacZ cells were T-cells containing cytosolic &gal activity; whereas >96% of the CTLLB cells were Thyl.2+ &gal-( Fig. 2A), >96% of the CTLL2IlacZ cells were Thyl.2+ ,&gal+ (Fig. 2B).
To exclude a possibility that retroviral integration of the foreign gene (la&) might change the growth characteristics of the CTLL2IlacZ cells, the proliferative response of CTLL2 and CTLL2/lacZ T-cells to IL-2 were compared. Importantly, the CTLL2IlacZ showed the same in uitro growth response to IL-2 as the unmodified CTLL2 T-cells (half-maximal  IL-2, compared with 0.84 f 0.37 for CTLL2/lacZ; p > 0.5), i.e. the modified cells maintained the responsiveness to IL-2 demonstrated by the parent cell line, offering the possibility that numbers of CTLL2/lacZ cells might be modulated in uiuo with exogenously administered IL-2.

Transplantation of Genetically Modified T-cells to the Respiratory Epithelial
Surface-Following the intratracheal transplant of 7.5 x 1O/g of body weight CTLL2IlacZ cells to the respiratory epithelial surface of the lung, no adverse effects attributable to the transplantation of the T-cells were noted. Flow cytometry evaluation of cells recovered from the respiratory ELF demonstrated that prior to transplantation, no Thyl.2+ @-gal+ T-cells were present (Fig. 2C). In contrast, 1 day following transplantation, large numbers of Thyl.2+ p-gal+ T-cells were evident (Fig. 2D). In the cells extracted from whole lung (which includes cells from the ELF and interstitium), pretransplantation showed primarily large, Thyl.2-cells (alveolar macrophages), with few Thyl.2+ @gal-T-cells and no @-gal+ T-cells (Fig. 2E). Similar to evaluation of the cells from the ELF, 1 day following transplantation, large numbers of Thyl.2+ @-gal+ cells were present (Fig. 2F). Histologic evaluation of lung sections from animals 1 day after transplantation with CTLL2/lacZ cells was consistent with findings from analysis of cells recovered from ELF and whole lung. Frozen lung sections stained with peroxidase-labeled avidin-biotin complex directed to the monoclonal mouse antibody Thyl.2 and stained with X-gal for the @-gal product revealed clumps of modified cells within the alveoli (i.e. on the respiratory epithelial surface) and within the alveolar interstitium (not shown).
Quantitative assessment of the Thyl.2+ T-cells present in the lung and spleen demonstrated that transplantation of CTLL2/lacZ cells to the respiratory epithelial surface resulted in the presence of large proportions of Thyl.2+ cells that were p-gal+ in respiratory ELF and in whole lung, but not in spleen (Fig. 3). As expected, the first day ( n = 4) after transplantation, most (95 f 3 % ) Thyl.2+ T-cells in respiratory ELF were p-gal+. A significant number of modified Tcells persisted in the lung for at least 1 week, as evidenced by the fact that at 7 days ( n = 7 ) , 33 f 8% of the Thyl.2+ cells in ELF were @-gal+. The same observation was made when the lung was minced and the T-cells from all compartments (ELF and the interstitial spaces) were evaluated. Interestingly, at 1 week, the proportions of Thyl.2+ cells that were P-gal+ were higher in whole lung than in ELF ( p < 0.01; n = 7 for ELF; n = 5 for whole lung) suggesting progressive movement of at least some of the modified T-cells from the epithelial surface into the parenchyma. Importantly, very small proportions of Thyl.2+ P-gal+ T-cells reached the systemic circulation as evidenced by the fact that there were never greater than 3 f 1% of Thyl.2+ cells in the spleen that were @-gal+ and, a t 1 week ( n = 12), no Thyl.2+ P-gal+ cells could be detected in the spleen.
Evaluation of cells from ELF at 2 weeks ( n = 2) demonstrated that approximately 30% of Thyl.2+ T-cells were @-gal+, suggesting that a reasonable number remained on the epithelial surface at this time. However, the lung tissue a t this time was likely repopulated by Thyl.2+ @-gal-cells, since Thyl.2+ (%gal+ cells were undetectable in the whole lung mince by 2 weeks ( n = 3 ) after transplantation.
Expansion of Thyl.2+ @-gal+ T-cells by Systemic Administration of IL-2"The decrease in the number of CTLL2/lacZ cells in the lung by 3 days post-transplantation, but with few modified T-cells in extrapulmonary lymph tissue, is consistent with the fact that CTLL2 T-cells require an exogenous source of IL-2 to remain viable. Accordant with the experience of other investigations (14,15,26), in vitro evaluation of CTLL2 cells cultured without IL-2 demonstrated progressive loss of the cells, such that, by 36 h, only 60% were viable and, by 48 h, only 31% were viable (as assessed by propidium iodide exclusion). Interestingly, administration of IL-2 to animals following intratracheal transplantation of CTLL2/ lac2 T-cells demonstrated an increase in the numbers of Thyl.2+ @-gal+ cells in the lung (Fig. 4). Animals treated with intraperitoneal human recombinant IL-2 post-transplantation showed no ill effects but had a higher number of CTLL2IlacZ cells in the lung. This was true when evaluated at the mRNA transcript level or by flow cytometry. Quantification of lac2 mRNA transcript levels in the lung demonstrated none were detectable pretransplantation, but abundant transcripts were detected 1 day after intratracheal transplantation of CTLLZIlacZ cells (Panel A). In contrast, under the conditions used, no lac2 transcripts were detectable by 3 days in animals receiving no IL-2. However, under the same conditions, in those animals receiving IL-2, lac2 transcripts were easily detectable at the same point in time following transplantation (compare lane 4 with lane 3 ) . The same pattern of results were observed using flow cytometry (Panel B ) . Three days after transplantation, the number of Thyl.2+ @-gal+ T-cells recovered from ELF was higher in animals treated with IL-2 ( n = 6) compared with that in untreated animals ( n = 5), but the difference was not statistically significant. However, there was a significant increase (4.6fold) of Thyl.2+ @-gal+ T-cells in whole lung of the IL-2treated animals ( p = 0.02; n = 8 with no IL-2; n = 8 with IL-2 treatment).

DISCUSSION
Genetically modified lymphocytes have the potential for treatment of a variety of human disorders, either because the modified lymphocytes have acquired an inherent lymphocyte feature that has relevance to the disease or by using the lymphocytes as a cell vehicle to synthesize and deliver a protein of therapeutic value (1)(2)(3)(4)(5)(6)(7)(8). A major hurdle for using modified lymphocyte gene therapy is how to achieve organ specificity in order to avoid an "unnecessary" burden of genetically modified cells at other sites. In the present study, using a murine model system, we have demonstrated that direct transplantation to an internal organ is a feasible approach. In this regard, genetically modified syngeneic T-cells transplanted to the respiratory epithelial surface remained in the lung for a t least 7 days. This was achieved without significant numbers of modified T-cells going beyond the lung. Further, the numbers of modified T-cells within the lung could be significantly increased by systemic administration of interleukin-2.
The T-lymphocyte is a convenient cell to use as a cell vehicle for gene therapy. In this regard, T-cells from animals and humans have been modified by gene transfer in uitro with a variety of human, animal, and bacterial genes including: a lantitrypsin, leukocyte adhesion glycoprotein LFA-1 (CD-18), IL-2, interleukin-4, interleukin-9, interferon-y, interferon-a?, growth hormone, adenosine deaminase, neomycin resistance, and lac2 (3, 6-8, 13, 26-36). Further, the modified cells can be expanded in large scale with the use of IL-2, and the Tcell populations can be evaluated prior to reinfusion for the presence and function of the new gene (3,6,7,30,32,36). This approach has been used in i n uiuo experimental animal studies to evaluate the contribution of IL-2, interleukin 4, and interleukin 9 to T-cell growth in uiuo (26,34,35), to track Tcells (3,6,30), and to evaluate the effect of lymphocytes modified with the adenosine deaminase gene (3,30). Exogenous gene transfer of human lymphocytes has been used in humans in uiuo to track tumor-infiltrating lymphocytes in individuals with melanoma and to treat adenosine deaminase deficiency (6,36).
Despite the accessibility and ease of growing T-cells in uitro, using T-cells as a target for gene transfer in humans has limited applicability unless the problem of targeting can be solved. First, because most diseases are localized, systemic administration of the modified T-cells without targeting has the problem of limiting therapeutic efficacy, as only a fraction of the modified T-cells will randomly reach the intended target. Although this might be solved by making each modified T-cell more "potent" (e.g. with more powerful promoters driving a gene for a therapeutic protein), there is the risk that the gene product will cause systemic toxicity. Second, although T-cells have receptors that are relevant to disease specificity (e.g. antigen receptors, homing receptors), to date, there is no clear evidence that systemic administration of a genetically modified subset of lymphocytes will "home" to a specific site.
The approach used in the present study may be one strategy to circumvent this targeting problem. It is particularly applicable to the lung, which is easily accessed via the respiratory tract. However, a similar approach might be used in other closed anatomic compartments, such as the central nervous system (via the cerebrospinal fluid), the bladder (retrograde via the urethra), and the gastrointestinal tract. Further, using selective catherization of localized arterial beds, modified Tcells could be transplanted to most organs.