Non-disrupted pieces of primary human lung tumor implanted into NOD-scid IL2Rγnull mice consistently result in successful xenografts in which tissue architecture, including tumor-associated leukocytes, stromal fibroblasts, and tumor cells are preserved for prolonged periods with limited host-vs-graft interference. Human CD45+ tumor-associated leukocytes within the xenograft are predominantly CD3+ T cells with fewer CD138+ plasma cells. The effector memory T cells that had been shown to be quiescent in human lung tumor microenvironments can be activated in situ as determined by the production of human IFN-γ in response to exogenous IL-12. Plasma cells remain functional as evidenced by production of human Ig. Significant levels of human IFN-γ and Ig were detected in sera from xenograft-bearing mice for up to 9 wk postengraftment. Tumor-associated T cells were found to migrate from the microenvironment of the xenograft to the lung, liver, and primarily the spleen. At 8 wk postengraftment, a significant portion of cells isolated from the mouse spleens were found to be human CD45+ cells. The majority of CD45+ cells were CD3+ and expressed a phenotype consistent with an effector memory T cell, consisting of CD4+ or CD8+ T cells that were CD45RO+, CD44+, CD62L, and CD25. Following adoptive transfer into non-tumor bearing NOD-scid IL2Rγnull mice, these human T cells were found to expand in the spleen, produce IFN-γ, and maintain an effector memory phenotype. We conclude that the NOD-scid IL2Rγnull tumor xenograft model provides an opportunity to study tumor and tumor-stromal cell interactions in situ for prolonged periods.

The first successful engraftment of human tumor tissue into mice homozygous for the scid (Prkdcscid) mutation was reported more than 20 years ago (1). Since this initial report, there have been several thousand reports on the engraftment into scid mice of a variety of different normal as well as neoplastic human cells and tissues. These studies have led to many important advances and insights with respect to human cancer, autoimmunity, and infectious diseases (2).

Engrafting non-disrupted pieces of human primary tumors into scid mice has served as a useful model for the evaluation of anti-cancer therapies. Such engraftment preserves the tumor microenvironment including tumor cells, infiltrating leukocytes, fibroblasts, extracellular matrix, and vasculature (3, 4). The engraftment of intact pieces of tumor instead of single cell suspensions has made it possible to study complex tumor-stroma interactions involved in the activation and migration of both tumor cells and tumor-associated cells (3, 4, 5, 6, 7, 8).

Experiments using the scid model have demonstrated that engraftment of a complete human tumor microenvironment can be preserved for a limited amount of time (3, 4, 9). Furthermore, some tumor-associated lymphocytes retain functionality, as indicated by the presence of human Ig in the mouse serum, and the production of human IFN-γ in response to intratumoral IL-12 cytokine stimulation (9, 10, 11, 12).

Although earlier studies have demonstrated the scid/human xenograft model to have many uses, limitations do exist. The CB17-scid mouse, which lacks functional B and T cells, maintains a functional innate immune system. Immediately following the implantation of human tissues, these mice mount a substantial host-vs-graft innate immune response (3, 4). This response involves murine granulocytes, macrophages, and NK cells, as well as various murine cytokines, all of which can persist for up to 2 wk postengraftment and have the potential to result in complete rejection of the xenograft and render it difficult to interpret results (3, 4, 13, 14).

To overcome the pitfalls related to the innate immune system in the scid mouse, a number of different strategies have evolved that selectively eliminate or block innate immune function. One example is to deplete the scid mouse of NK cells with a polyclonal Ab to asialo GM-1. However, its effects are transient and NK cell numbers begin to recover within 9 days (3, 4, 15). In our laboratory and others, NK cells have been depleted for up to 6 wk using the mAb, TMβ-1, directed against the β-chain of the IL-2 receptor (16, 17). Xenografts in scid mice depleted of NK cells have improved tumor engraftment and permitted lengthier studies (3, 4, 10, 11, 12). However, as with any Ab treatment, results are dependent upon the route of administration, dose, and schedule of treatment. In addition, treatment with Abs often results in only a partial depletion of NK cells and other host factors that contribute to the innate response to xenografts (3, 4). Currently, there is no standardized protocol for the depletion of the scid mouse innate response that has provided consistently reproducible results of long term tumor engraftment. Methods vary between laboratories making it difficult to interpret and duplicate results. As previously suggested (3, 4), a standardized protocol should be developed in which optimum depletion of murine innate function is achieved for the purpose of eliminating inconsistencies.

An alternative method for depletion of NK activity is the further genetic alteration of the scid mouse. The first step toward this goal was the development of the NOD/Lt-scid (NOD-scid) mouse. These mice were shown to have decreased NK cell function, a lack of hemolytic complement, and functionally immature macrophages (18, 19, 20, 21). However, NOD-scid mice have a relatively short life span due to the development of early-onset thymic lymphomas, as well as residual NK cell activity (18). These shortcomings were eliminated in the development of a NOD-scid mouse with a null mutation of the IL-2 receptor common γ-chain (18, 22, 23, 24). This mutation results in a defect of the high affinity receptors for IL-2, IL-4, IL-7, IL-9, IL-15, and IL-21 (18, 25, 26) that blocks the development of NK cells and further impairs innate immunity (18, 27). Additionally, it was found that NOD-scid IL2Rγnull mice are less likely to develop thymic lymphomas and are long lived (18). Recent experiments have shown that NOD-scid IL2Rγnull mice are highly supportive of the engraftment of normal human leukocytes. NOD-scid IL2Rγnull mice have increased percentages of human CD45+ cells following implantation of cytokine mobilized hematopoietic stem cells (18) and sustained development of human hemato-lymphopoiesis after the injection of CD34+ umbilical cord blood cells (18, 28).

We report here the use of the NOD-scid IL2Rγnull mouse for the engraftment of non-disrupted pieces of human primary lung tumor. Using the NOD-scid IL2Rγnull mouse model it is possible to successfully engraft non-disrupted pieces of human lung tumor and maintain the tumor microenvironment in the absence of additional cell depletion. These xenografts survived for up to 9 wk postengraftment. Tumor and tumor-associated stromal cells remained viable and functional in the absence of high levels of mouse cell infiltrates into the engrafted tissue. Effector memory T cells present within the microenvironment of the xenografts were found to produce IFN-γ in response to exogenous IL-12, and a portion of these T cells migrated out of the xenografts to mouse spleen, lung, and liver. It was also determined that the tumor-associated T cells recovered from the spleens of mice bearing tumor xenografts could be maintained and expanded after adoptive transfer into tumor-free NOD-scid IL2Rγnull mice.

Primary non-small cell lung tumor biopsy specimens were received from Roswell Park Cancer Institute Tissue Procurement Facility, Department of Pathology at Kenmore Mercy, or Millard Fillmore-Gates Circle Hospitals. All specimens were obtained under sterile conditions and using Institutional Review Board approved protocols. Histological diagnosis for each specimen was obtained anonymously at a later date.

As previously described (14), necrotic tissue was removed from the tumor specimens and sections (3–5 mm on a side) were cut resulting in cubic pieces of non-disrupted tumor. CB17-scid mice were obtained from the breeding colony at the State University of New York (SUNY) at Buffalo and NOD-scid IL2rgtmWjl/J (NOD-scid IL-2Rγnull) mice were obtained from The Jackson Laboratory. All mice were housed in specific pathogen-free conditions at SUNY. Before implantation, mice were anesthetized with Avertin (10 mg/mouse; Sigma-Aldrich) and the CB17-scid mice were treated with one i.p. injection of TMβ-1 (mAb to murine IL-2R β-chain) (16, 17). A small ventral midline incision was made just caudal to the xiphoid process, and a s.c. pocket was created. One piece of tumor was implanted into the pocket and the incision closed using Nexaband Liquid (Burns Veterinary Supply). All animal experiments were approved by the Institutional Animal Care and Use Committee at SUNY.

Recombinant human IL-12 (Wyeth Research) was incorporated into biodegradable microspheres with BSA and polylactic acid at a concentration of 0.1 mg of recombinant human IL-12/100 mg of polymer. Mice bearing xenografts were randomly divided into control and treatment groups 7–10 days after surgical implantation of fresh tumor tissue. All mice were treated with a single injection of IL-12 loaded microspheres (10 μg/mouse) or an equivalent weight of BSA microspheres.

Blood samples were collected via a tail clipping or retro-orbital sinus bleed. Sera or plasma were collected by microcentrifugation at 14,000 rpm for 10 min and stored at −20°C until use.

A sandwich ELISA for the detection of human IFN-γ in mouse serum was performed as previously described (29, 30). Briefly, 96-well plates were coated using a monoclonal anti-human IFN-γ Ab (M-700-A; Endogen). Mouse sera were added to the plate with biotinylated monoclonal anti-human IFN-γ (M-701-B; Endogen). Positive binding was detected using streptavidin-conjugated HRP (Sigma-Aldrich), peroxide, and 3,3′, 5,5′-tetramethylbenzidine (Kirkegaard & Perry Laboratories). Results were measured on a Bio-Tek Instruments automated microplate reader at OD450-540 and analyzed by comparison with a recombinant human IFNγ standard using SigmaPlot software.

An ELISA for the detection of human Ig in mouse serum was performed as previously described (9, 31, 32). Briefly, 96-well plates were coated using rabbit anti-human Ig (Accurate Chemical & Scientific), and incubated overnight at 4°C. The plates were washed. Bound human Ig was detected using rabbit anti-human Ig-HRP and a substrate mixture containing peroxide and o-phenylenediamine (Sigma-Aldrich). Results were measured on a Bio-Tek Instruments automated microplate reader at OD490 and analyzed by comparison with a human Ig standard using SigmaPlot software.

Human leukocytes were analyzed phenotypically using four-color flow cytometry. Single cell suspensions were derived from surgically excised xenografts, spleens, lungs, peripheral blood, and bone marrow. Tissues were mechanically disrupted using a Teflon policeman to gently force cells through a size 50 metal mesh. Bone marrow was obtained by removing and cutting both ends of the femurs and flushing the marrow plugs with PBS using a needle and syringe. Tissues were pooled according to the original tumor, tissue type, and treatment group. For the phenotyping of the human CD3+ T cell population, single cell suspensions were either treated with a RBC lysis buffer (eBioscience) or centrifuged on a Ficoll-Paque PLUS gradient (GE Healthcare).

The cell suspensions were stained with multiple Ab panels. Each panel contained four Abs bearing a different fluorochrome. The human mAbs included in the panels were anti-CD3, anti-CD8, anti-CD4, anti-CD25, anti-CD45, anti-CD45RA, anti-CD45RO, anti-CD69, anti-CD62L, anti-CD11a, anti-CXCR3, anti-CD44 (BD Pharmingen), and anti-FoxP3 (eBioscience). The cells stained for intracellular FoxP3 were fixed and permeabilized using the FoxP3 staining buffer set (eBioscience). Data were collected on a FACSCalibur flow cytometer (BD Biosciences) and analyzed using WinList software (Verity). One hundred thousand events or more were collected for each panel.

H&E staining was performed by the SUNY Histology Service Laboratory where fresh surgical sections of the original tumor, xenografts, and mouse tissues were fixed in 10% neutral-buffered formalin and processed for paraffin embedding. Anti-human Abs to the following markers were used for immunohistochemical staining of xenografts and mouse tissues: anti-CD3, anti-CD138 (DakoCytomation), anti-HLA-A (A-18; Santa Cruz Biotechnology), anti-IgG + A + M-HRP (Zymed Laboratories), and anti-mouse CD45 (BD Pharmingen). Staining was performed according to the manufacturer’s instructions (DakoCytomation). The AE1/AE3 cytokeratin (Thermo Shandon) and Ki67 (NeoMarker) staining was performed by the pathology core facility at Roswell Park Cancer Institute. The Ki67 proliferation index was calculated by counting the total number of positively stained cells and dividing this by the total number of cells in that field. Six different fields for each slide (both the original tumor and xenografts) were counted, and an average and SD were calculated for each slide. H&E and immunohistochemistry images were taken with a Zeiss Axioimager using Axiovision Rel 4.6 Software.

Cells derived from the spleens of xenograft-bearing mice (obtained from a Ficoll-Paque Plus gradient) were washed in serum-free medium and resuspensed in sterile 0.9% saline to a concentration of 5 × 106 cells/100 μl. Approximately 5 × 106 cells were injected into the lateral tail vein of each mouse.

To determine whether it is possible to engraft a human tumor microenvironment into NOD-scid IL2Rγnull mice, non-disrupted pieces of primary non-small cell lung tumors were implanted subcutaneously into these mice. As a comparison to our current model of tumor engraftment, CB17-scid mice were also implanted with non-disrupted pieces of tumor. Animals were sacrificed and xenografts were removed between 7 and 9 wk postimplantation. The xenografts of the NOD-scid IL2Rγnull mice showed histological features that were present within the tumor microenvironment of the original biopsy (Fig. 1, A and D). Tumor matched CB17-scid mouse xenografts were not structurally maintained and appeared to contain mostly fibrotic structures and host granulocytes 9 wk after implantation (Fig. 1 J).

FIGURE 1.

H&E and immunohistochemistry sections of a primary non-small cell lung tumor and corresponding xenografts in NOD-scid IL2Rγnull and CB17-scid mice 9 wk postengraftment. H&E sections of a primary lung adenosquamous cell carcinoma (A) before engraftment and corresponding xenografts removed 9 wk postengraftment from NOD-scid IL2Rγnull (D and G) and CB17-scid mice (J). The neoplastic cells in the original tumor are universally positive for cytokeratin staining (B). The corresponding NOD-scid IL2Rγnull tumor xenograft demonstrates retention of the cytokeratin immunoreactivity by the tumor cells (E). The Ki67 staining was positive in both the original tumor (C) and xenograft (F). Images were taken at ×400 magnification. As opposed to the CD17-scid mouse xenograft (K) the NOD-scid IL2Rγnull mouse xenograft (H) was highly infiltrated with HLA-A positive cells. There were also very few murine CD45 positive cells in the NOD-scid IL2Rγnull mouse xenograft (I) when compared with the CB17-scid mouse xenograft (L), which was rapidly infiltrated with murine CD45+ cells. Images were taken at ×100 magnification.

FIGURE 1.

H&E and immunohistochemistry sections of a primary non-small cell lung tumor and corresponding xenografts in NOD-scid IL2Rγnull and CB17-scid mice 9 wk postengraftment. H&E sections of a primary lung adenosquamous cell carcinoma (A) before engraftment and corresponding xenografts removed 9 wk postengraftment from NOD-scid IL2Rγnull (D and G) and CB17-scid mice (J). The neoplastic cells in the original tumor are universally positive for cytokeratin staining (B). The corresponding NOD-scid IL2Rγnull tumor xenograft demonstrates retention of the cytokeratin immunoreactivity by the tumor cells (E). The Ki67 staining was positive in both the original tumor (C) and xenograft (F). Images were taken at ×400 magnification. As opposed to the CD17-scid mouse xenograft (K) the NOD-scid IL2Rγnull mouse xenograft (H) was highly infiltrated with HLA-A positive cells. There were also very few murine CD45 positive cells in the NOD-scid IL2Rγnull mouse xenograft (I) when compared with the CB17-scid mouse xenograft (L), which was rapidly infiltrated with murine CD45+ cells. Images were taken at ×100 magnification.

Close modal

To determine whether the tumor cells in the xenografts retained their neoplastic potential after engraftment into NOD-scid IL2Rγnull mice, AE1/AE3 cytokeratin and Ki67 immunohistochemistry staining was performed on both the original tumor and corresponding xenografts. As can be seen in Fig. 1,B, the neoplastic cells in the original tumor are universally positive for cytokeratin. At 9 wk postimplantation, the xenografts isolated from NOD-scid IL2Rγnull mice demonstrated retention of the cytokeratin immunoreactivity in the tumor cells and an absence of this staining in the surrounding inflammatory milieu. Furthermore, Ki67 staining, which predicts the proliferation or growth fraction of the tumor, was positive in both the original tumor and xenografts (Fig. 1, C and F, respectively). Analysis of the Ki67 proliferation index revealed an average of 41.6 ± 3.8% in the original tumor and 39.0 ± 4.4% in the corresponding xenograft. Therefore, as in the original tumor, the tumor cells in the xenograft appear to be neoplastic.

The same xenografts were then subjected to various other immunohistochemical stains to characterize differences (if any) between the CB17-scid and NOD-scid IL2Rγnull xenografts. Staining for HLA-A expression revealed that the majority of the xenograft tissue from the NOD-scid IL2Rγnull mice was positive for HLA-A at 9 wk postengraftment (Fig. 1,H), whereas significant portions of xenografts in the CB17-scid mice were found to be HLA-A negative (Fig. 1,K), and therefore presumably mouse tissues. Murine CD45 staining revealed a minimal invasion of host leukocytes into the NOD-scid IL2Rγnull mouse xenografts, contrasting with dense accumulations of host leukocytes in the CB17-scid mice (Fig. 1, I and L, respectively). The presence of host leukocytes has previously been shown following long-term engraftment of human tumors in CB17-scid mice (3, 4, 13, 14, 33). Together, these results demonstrate that the NOD-scid IL2Rγnull mouse is capable of engrafting and maintaining a structurally complete human tumor microenvironment for periods of up to 9 wk in which the majority of tissues remaining in the xenograft express human HLA-A and relatively negligible murine leukocyte infiltration.

After successful engraftment was achieved in the NOD-scid IL2Rγnull mouse, we wanted to determine whether the inflammatory cells present within the tumor xenograft microenvironment were functional. Previous short-term (2–3 wk postengraftment) studies of tumor xenografts in CB17-scid mice established that IL-12 leads to the activation of quiescent human CD4+ and CD8+ T cells in the microenvironment of the xenografts and induces secretion of IFN-γ, which peaked 5 days post-treatment and rapidly returned to baseline levels (11, 12). This early response is also observed in the NOD-scid IL2Rγnull mouse model, as demonstrated by the presence of human IFN-γ occurring at day 5 postintratumoral treatment with IL-12 loaded microspheres. No IFN-γ was detected in the control mice at 5 days post-treatment (Fig. 2,A). In contrast to the CB17-scid mouse model (11, 12), elevated levels of IFN-γ were still detected in the serum of NOD-scid IL2Rγnull mice at 4 wk post IL-12 microsphere treatment and were sustained in these mice until termination of the experiment at 8 wk post-treatment (Fig. 2,A). Although the absolute amount of IFN-γ varies between mice within the same treatment group or from tumor to tumor, the trend of IFN-γ production is consistent between experiments. This demonstrates that inflammatory cells within the tumor bearing mice persist and are functional for at least 8 wk post-treatment. Four weeks postengraftment, low levels of IFN-γ were detected in the sera of BSA control microsphere-treated xenografts (Fig. 2 A). These results suggest that during a prolonged engraftment period, tumor-derived lymphocytes continue to secrete IFN-γ in response to IL-12 treatment up to 8 wk post-treatment. In addition, tumor-derived lymphocytes secrete low levels of IFN-γ in the absence of exogenous IL-12.

FIGURE 2.

Human IFN-γ and Ig levels were detected in the serum of NOD-scid IL2Rγnull mice implanted with a primary human lung tumor at various time points following treatment. Human IFN-γ was sustained in response to IL-12 stimulation for the duration of the experiment (A). IFN-γ was detected in the serum of control mice (BSA) at 4 wk post-treatment. Both the BSA- and IL-12-treated mice produced human Ig for the entire length of the experiment (B). Data are given from one representative experiment (n = 4).

FIGURE 2.

Human IFN-γ and Ig levels were detected in the serum of NOD-scid IL2Rγnull mice implanted with a primary human lung tumor at various time points following treatment. Human IFN-γ was sustained in response to IL-12 stimulation for the duration of the experiment (A). IFN-γ was detected in the serum of control mice (BSA) at 4 wk post-treatment. Both the BSA- and IL-12-treated mice produced human Ig for the entire length of the experiment (B). Data are given from one representative experiment (n = 4).

Close modal

We next examined the functionality of engrafted B-lymphocytes and/or plasma cells through measurements of mouse serum human Ig. Measurable levels of human Ig were present in the mouse sera as early as 5 days after IL-12 treatment (Fig. 2 B), and increased levels of Ig were seen at 4 and 8 wk post-treatment. We conclude that human tumor-associated T cells and B/plasma cells remain functional for at least 9 wk following tumor implantation into NOD-scid IL2Rγnull mice.

To determine whether the tumor-associated lymphocytes remained in the xenograft or migrated to secondary lymphoid and non-lymphoid tissues, peripheral blood, spleen, lung, bone marrow, and xenografts were collected from NOD-scid IL2Rγnull mice 7–9 wk following tumor engraftment. Four-color flow cytometry was performed to identify and phenotype the human leukocytes that may be present in various mouse organs and tissues. It was observed that 1–3% of the total cells in the lung, bone marrow, and xenografts were human CD45+ leukocytes. No CD45+ cells were detected in the peripheral blood 9 wk after engraftment (data not shown). After mechanically disrupting the spleens and lysing the RBC, ∼32% of the splenocytes isolated from the xenograft-bearing mice were human CD45+ cells (Fig. 3,A). The human CD45+ cells isolated from the spleen were also found to be predominately CD3+ T cells (Fig. 3 B). Therefore, while some human leukocytes remained within the xenograft, a larger number of these leukocytes migrated to mouse tissues, primarily the spleen. The high percentage of human cells outside of the xenograft suggests that the tumor-derived leukocytes expanded either prior or subsequent to leaving the tumor xenografts.

FIGURE 3.

The majority of CD45+ cells found within the spleen of tumor bearing NOD-scid IL2Rγnull mice are human CD3+ T cells. Mice were sacrificed 9 wk postengraftment. The spleens were removed, pooled, mechanically disrupted, and treated with RBC lysis buffer. Approximately 32% of the cells in the spleens were human CD45+ cells (A). After gating on the CD45+ population, the majority of the cells were CD3+ T lymphocytes (96%) (B). The solid black line represents the autofluorescence, and the gray shaded area represents the stained population of cells. At least 100,000 events were collected.

FIGURE 3.

The majority of CD45+ cells found within the spleen of tumor bearing NOD-scid IL2Rγnull mice are human CD3+ T cells. Mice were sacrificed 9 wk postengraftment. The spleens were removed, pooled, mechanically disrupted, and treated with RBC lysis buffer. Approximately 32% of the cells in the spleens were human CD45+ cells (A). After gating on the CD45+ population, the majority of the cells were CD3+ T lymphocytes (96%) (B). The solid black line represents the autofluorescence, and the gray shaded area represents the stained population of cells. At least 100,000 events were collected.

Close modal

Within the spleen, the majority of human T cells were found to have a 2:1 ratio of CD4+:CD8+, and consisted primarily of CD45RO+, CD11a+, CXCR3+, CD44+, CD69, CD62L, CD25, and FoxP3 cells (Fig. 4). This is indicative of an effector memory phenotype. There was also a small, but distinct, percentage of CD25+FoxP3+ regulatory T cells (0.74%) in the spleens 9 wk post tumor engraftment.

FIGURE 4.

The phenotype of human CD3+ T cells within the spleens of NOD-scid IL2Rγnull mice engrafted with a human lung tumor are indicative of effector memory T cells. These cells had approximately a ratio of 2:1 for CD4+:CD8+ (A). The majority of the T cells were found to be CD45RO+ (B), CD69 and CD62L (C), CXCR3+ and CD11a+ (D), CD25 and FoxP3 (E), and CD44+ (F). All populations are gated on human CD3+ T cells.

FIGURE 4.

The phenotype of human CD3+ T cells within the spleens of NOD-scid IL2Rγnull mice engrafted with a human lung tumor are indicative of effector memory T cells. These cells had approximately a ratio of 2:1 for CD4+:CD8+ (A). The majority of the T cells were found to be CD45RO+ (B), CD69 and CD62L (C), CXCR3+ and CD11a+ (D), CD25 and FoxP3 (E), and CD44+ (F). All populations are gated on human CD3+ T cells.

Close modal

The presence of human leukocytes in mouse tissues following IL-12 treatment of xenografts was confirmed using immunohistochemical staining of tissue sections. The xenografts, mouse spleen, lung, liver, kidney, and intestine were stained using anti-human CD3 and CD138 to identify T cells and plasma cells, respectively. CD3+ T cells (Fig. 5,A) and CD138+ cells (Fig. 5,G) with a plasmacytoid-like morphology were observed in discrete foci within the xenograft. CD3+ human T cells were shown to be present in the spleen (Fig. 5,B), lung (Fig. 5,C), liver (Fig. 5,D), kidney (Fig. 5,E), and lamina propria of the small intestine (Fig. 5,F) of tumor xenograft-bearing mice 6 wk postengraftment. CD138+ plasma cells were found in the spleen (Fig. 5,H) and lung (Fig. 5 I) but were noticeably less than that observed in the tumor xenografts.

FIGURE 5.

Tumor-associated T cells and plasma cells relocate from the xenograft to other mouse tissues and organs. Mice were implanted with a human primary lung tumor, and at 5 wk post IL-12 treatment (6 wk postengraftment) mice were sacrificed and various tissues removed. Immunohistochemical staining analysis revealed the presence of CD3+ T cells in the xenograft (A), spleen (B), and lung (C). Small discrete foci of CD3+ T cells were shown to be present in the liver (D), kidney (E), and lamina propria of the intestine (F) of the xenograft-bearing mice. Immunohistochemical staining with human αCD138 revealed plasma cells in the xenograft (G) with fewer plasma cells seen in the spleen (H) and lung (I). All images were taken at ×100 magnification.

FIGURE 5.

Tumor-associated T cells and plasma cells relocate from the xenograft to other mouse tissues and organs. Mice were implanted with a human primary lung tumor, and at 5 wk post IL-12 treatment (6 wk postengraftment) mice were sacrificed and various tissues removed. Immunohistochemical staining analysis revealed the presence of CD3+ T cells in the xenograft (A), spleen (B), and lung (C). Small discrete foci of CD3+ T cells were shown to be present in the liver (D), kidney (E), and lamina propria of the intestine (F) of the xenograft-bearing mice. Immunohistochemical staining with human αCD138 revealed plasma cells in the xenograft (G) with fewer plasma cells seen in the spleen (H) and lung (I). All images were taken at ×100 magnification.

Close modal

The spleens of tumor bearing NOD-scid IL2Rγnull mice were removed 6 wk postengraftment, and cell suspensions were centrifuged on a Ficoll-Paque Plus gradient. After separation, 5 × 106 of these cells (which contained 96% human CD3+ T cells) were adoptively transferred into non-tumor bearing NOD-scid IL2Rγnull mice. Mice were monitored for human T and B cell activity by assaying the mouse sera for human IFN-γ and Ig at biweekly intervals. Elevated levels of human IFN-γ but not Ig were found in the sera at all assayed time points. This suggests that T cells, but not a significant population of B cells or plasma cells, were present and functional throughout the 6 wk period of observation (Fig. 6, A and B).

FIGURE 6.

T cells isolated from the spleens of NOD-scid IL2Rγnull mice adoptively transferred with tumor-associated T cells derived from the spleens of tumor xenograft-bearing mice have a phenotype indicative of effector memory T cells. 5.0 × 106 cells isolated from the spleens of tumor xenograft-bearing mice were adoptively transferred into NOD-scid IL2Rγnull mice. The mice were bled at 2, 4, and 6 wk after the adoptive transfer. The mice were sacrificed 6 wk after adoptive transfer, the spleens were removed, mechanically disrupted, and isolated on a Ficoll-Paque Plus gradient. Serum samples demonstrated an increase in human IFN-γ throughout the duration of the experiment (A). No human Ig was detected at 2, 4, and 6 wk post adoptive transfer (B). Four-color flow cytometric analysis identified the T cells isolated from the spleens of the adoptively transferred mice to have an effector memory phenotype. The majority of CD3+ T cells were found to be CD4+ or CD8+ T cells (C), CD45RO+ and CD45RA (D), CD69 and CD62L (E), CXCR3+and CD11a+ (F), CD25 (G), and CD44+ (H). All populations were gated on human CD3+ T cells. Data are given from one representative experiment (n = 2).

FIGURE 6.

T cells isolated from the spleens of NOD-scid IL2Rγnull mice adoptively transferred with tumor-associated T cells derived from the spleens of tumor xenograft-bearing mice have a phenotype indicative of effector memory T cells. 5.0 × 106 cells isolated from the spleens of tumor xenograft-bearing mice were adoptively transferred into NOD-scid IL2Rγnull mice. The mice were bled at 2, 4, and 6 wk after the adoptive transfer. The mice were sacrificed 6 wk after adoptive transfer, the spleens were removed, mechanically disrupted, and isolated on a Ficoll-Paque Plus gradient. Serum samples demonstrated an increase in human IFN-γ throughout the duration of the experiment (A). No human Ig was detected at 2, 4, and 6 wk post adoptive transfer (B). Four-color flow cytometric analysis identified the T cells isolated from the spleens of the adoptively transferred mice to have an effector memory phenotype. The majority of CD3+ T cells were found to be CD4+ or CD8+ T cells (C), CD45RO+ and CD45RA (D), CD69 and CD62L (E), CXCR3+and CD11a+ (F), CD25 (G), and CD44+ (H). All populations were gated on human CD3+ T cells. Data are given from one representative experiment (n = 2).

Close modal

Mice were sacrificed 6 wk after adoptive transfer and the spleens of the recipient mice were removed. The pooled spleens were mechanically disrupted and isolated on a Ficoll-Paque Plus gradient. Approximately 3.4 × 107 cells were recovered from the spleens of three mice that received an injection of 5 × 106 cells/mouse. This represents an approximate 2.25-fold expansion of these cells. Given the inefficiency of isolating and recovering cells from the spleen and the expected failure of all of the injected cells to reach the spleen, this is likely a significant underestimate of the actual increase of the cells following their transfer into non-tumor bearing mice. The presence of human T cells in the recipient mice was demonstrated in situ with anti-human CD3 immunohistochemical staining of the mouse spleen, lung, and liver (data not shown). This demonstrated large numbers of repopulating cells in patterns indistinguishable from those of tumor engrafted mice and confirmed the expansion of these cells in the recipient mice. The cells recovered from mice adoptively transferred with tumor-derived lymphocytes were subjected to flow cytometric analysis. Ninety-six percent of the cells were CD3+ T cells (data not shown). The T cell population was found to contain both CD3+CD4+ and CD3+CD8+ T cells. The majority of these cells had an effector memory phenotype of CD45RO+, CD11a+, CXCR3+, CD44+, CD69, CD62L, and CD25 (Fig. 6, CH). These results demonstrate that human lymphocytes originating from an engrafted tumor are able to survive and be adoptively transferred into non-tumor bearing NOD-scid IL2Rγnull mice. Once transferred, these human CD45+ cells expand, repopulate the spleen, and remain phenotypically indistinguishable from the T cells that were originally isolated from the tumor bearing spleen.

We report here that human non-small cell lung tumor microenvironments can be successfully implanted into NOD-scid IL2Rγnull mice, and the resultant xenografts are maintained both structurally and functionally for up to 9 wk. This new model has made it possible to study tumor and stromal cell interactions for extended periods in situ with minimal infiltration of murine host immunocompetent cells into the human xenograft. This model can be used to enhance and standardize the protocol for the implantation of human tumors in which there is a sustained presence of a functional tumor microenvironment.

It was found that some of the tumor-derived human leukocytes migrate from the tumor microenvironment to several different mouse organs including the spleen, lung, liver, kidney, and intestine when implanted into a NOD-scid IL2Rγnull mouse. These results differ from previous reports that tumor-associated leukocytes remain in the tumor xenograft when implanted into CB17-scid mice (3, 4, 11).

Another significant difference between the CB17-scid and NOD-scid IL2Rγnull mouse is that the tumor-associated T cells derived from NOD-scid IL2Rγnull mice are sustained and expand. These T cells were found to be mostly positive for CD4 or CD8, have an effector memory phenotype, and to be mostly negative for CD25+FoxP3+ regulatory T cells (<1%).

Several possible factors could be contributing to the sustained presence and expansion of the tumor-associated T cells that are observed in the NOD-scid IL2Rγnull mouse model. One possibility is that knocking out the common γ-chain of the IL-2 receptor may result in the production and accumulation of enhanced amounts of murine cytokines that are produced as a result of the absence of corresponding cytokine receptors. Murine IL-7 (mIL-7)4 requires that the common γ-chain of the IL-7 receptor be intact for high affinity binding to its receptor. Therefore, if the common γ-chain is absent, mIL-7 is expected to be elevated due to positive feedback. Human IL-7 has been found to be responsible for the survival of memory T cells (34), and mIL-7 has been shown to have reactivity on human lymphocytes (35, 36). If mIL-7 is present and/or elevated in the NOD-scid IL2Rγnull mouse, this could contribute to the sustained presence of tumor-associated T cells observed in this xenograft model. We have measured serum mIL-7 in several implanted mice and found undetectable levels (unpublished data) suggesting that mIL-7 is not likely contributing to the prolonged survival and expansion of tumor-derived T cells in the NOD-scid IL2Rγnull mice. The complete absence of an adaptive immune response and the greatly reduced innate immune response of NOD-scid IL2Rγnull is likely to contribute most significantly to the prolonged survival of the xenografts and to the sustained presence and expansion of the tumor-associated T cells.

Despite the successful engraftment of human tumors, the prolonged presence of tumor-associated T cells, B cells, and plasma cells pose a potential limitation to this xenograft model. We and others have previously noted that CB17-scid mice engrafted with human PBMCs eventually develop xenoreactive Abs and T lymphocytes (31, 32, 37). Recent studies in NOD-scid IL2Rγnull mice have observed symptoms of xenoreactivity as early as 30 days after the injection of human PBMC (38). In contrast, histological examination of the spleen, lung, liver, and skin in xenograft implanted NOD-scid IL2Rγnull mice showed no evidence of xenoreactive graft-vs-host disease (XGVHD). Moreover, no clinical or histological evidence of xenoreactivity have been observed in NOD-scid IL2Rγnull mice following the injection of human T cells derived from the spleens of xenograft-bearing mice. The mice were observed for up to 45 days with no evidence of weight loss, decreases in hematocrit, or histological evidence of XGVHD (data not shown). It has been established that tumor-associated cells do not induce XGVHD in CB17-scid mice following the engraftment of human lung tumors (9). However, in view of the prolonged survival of the tumor-associated T cells in the NOD-scid IL2Rγnull mice and continued production of IFN-γ following their transfer into mice, it will be important to further address the issue of XGVHD. Because the T cells injected into and recovered from the spleens of the NOD-scid IL2Rγnull mice have the same phenotype (i.e., effector memory T cell) as the T cells found in the original tumor, at least two alternative possibilities could explain the continuous activation of these tumor-associated effector memory T cells. First, that the effector memory T cells remain activated due to the absence of normally present regulatory mechanisms. Consistent with this possibility is that the regulatory T cells which have been observed in the original tumor are mostly absent from the tumor-associated T cells found in the spleens of xenograft-bearing mice. The second possibility is that the APCs along with Ags from the original tumor microenvironment may be activating the transferred T cells. A small but distinct population of cells with the phenotype of an APC (i.e., positive for HLA-DR and positive for either CD11c, CD19, or CD14) has been identified in the otherwise predominant T cell population derived from the spleens of the xenograft-bearing mice (data not shown).

The ability to establish and maintain human tumor microenvironments by surgically implanting non-disrupted pieces of lung tumor tissue into NOD-scid IL2Rγnull mice provides a unique model with which to study complex interactions in vivo that occur between tumor cells and stroma, which includes inflammatory leukocytes, fibroblasts, and the extracellular matrix. This model is expected to be particularly useful in assessing the efficacy of novel therapeutic approaches to cancer that target the tumor, the tumor stroma, or both the normal and neoplastic cells within the tumor microenvironment.

The expansion of tumor-associated leukocyte subsets (in this case CD4+ and CD8+ T cells with an effector memory phenotype), and our ability to recover these cells from human tumor bearing mice and adoptively transfer them into other NOD-scid IL2Rγnull mice may provide a valuable resource in the studies of basic tumor-associated T cell biology. However, it will first be important to establish that the tumor-derived cells are not being driven by xenoantigens. It is expected that such studies may help to explain why the tumor-associated T cells fail to control tumor progression. Previous studies at the single-cell level have established that effector memory T cells within the microenvironment of human non-small cell lung tumors are hyporesponsive to activation via the TCR and CD28 (39). This T cell quiescence is at least partially mediated by a membrane bound TGF-β (40) that is produced by tumor-associated myofibroblasts (41). Our preliminary studies suggest that the T cells fail to be activated due to a regulatory checkpoint in the TCR signaling cascade that is located upstream of phospholipase C-γ (unpublished data). A more precise localization and characterization of the putative regulatory signaling checkpoint may now be achieved due to our ability to generate large numbers of the tumor-associated T cells. It should also be possible with this new humanized scid model to determine what effect the tumor stroma has upon the regulation of the tumor-associated T cell function and its ability to recognize and kill tumor cells in situ.

We thank Jennifer Barnas, Dr. Stephen Brooks, and Liza Pope for assistance during tissue specimen collections. Special thanks also to Jenni Loyall and Robert Parsons for outstanding technical assistance, Dr. Lori Broderick for performing the mouse cytokine ELISA, and Dr. Todd Demmy and tissue procurement at Roswell Park Cancer Institute for providing one of the patient tumor samples. We also gratefully acknowledge Craig Abelson for critically reviewing the manuscript.

Dr. Richard B. Bankert is the president of Therapyx Inc.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported in part by U.S. Public Health Service Grants, National Institutes of Health Grant R01-CA10897 (to R.B.B.), the John R. Oishei Foundation, National Institutes of Health Research Training Grant T32 AI1007614-07 (to M.R.S.-A.), and the Jackson Lab Cancer Core Grant CA34196 (to L.D.S.)

4

Abbreviations used in this paper: mIL-7, murine IL-7; XGVHD, xenoreactive graft vs host disease.

1
Reddy, S., D. Piccione, H. Takita, R. B. Bankert.
1987
. Human lung tumor growth established in the lung and subcutaneous tissue of mice with severe combined immunodeficiency.
Cancer Res.
47
:
2456
-2460.
2
Shultz, L. D., F. Ishikawa, D. L. Greiner.
2007
. Humanized mice in translational biomedical research.
Nat. Rev. Immunol.
7
:
118
-130.
3
Bankert, R. B., N. K. Egilmez, S. D. Hess.
2001
. Human-SCID mouse chimeric models for the evaluation of anti-cancer therapies.
Trends Immunol.
22
:
386
-393.
4
Bankert, R. B., S. D. Hess, N. K. Egilmez.
2002
. SCID mouse models to study human cancer pathogenesis and approaches to therapy: potential, limitations, and future directions.
Front. Biosci.
7
:
c44
-c62.
5
Dustin, M. L., A. R. de Fougerolles.
2001
. Reprogramming T cells: the role of extracellular matrix in coordination of T cell activation and migration.
Curr. Opin. Immunol.
13
:
286
-290.
6
Ibe, S., Z. Qin, T. Schuler, S. Preiss, T. Blankenstein.
2001
. Tumor rejection by disturbing tumor stroma cell interactions.
J. Exp. Med.
194
:
1549
-1559.
7
McCawley, L. J., L. M. Matrisian.
2001
. Tumor progression: defining the soil round the tumor seed.
Curr. Biol.
11
:
R25
-R27.
8
Tlsty, T. D..
2001
. Stromal cells can contribute oncogenic signals.
Semin. Cancer Biol.
11
:
97
-104.
9
Williams, S. S., F. A. Chen, H. Kida, S. Yokata, K. Miya, M. Kato, M. P. Barcos, H. Q. Wang, T. Alosco, T. Umemoto, et al
1996
. Engraftment of human tumor-infiltrating lymphocytes and the production of anti-tumor antibodies in SCID mice.
J. Immunol.
156
:
1908
-1915.
10
Sugiyama, Y., M. Kato, F. A. Chen, S. S. Williams, Y. Kawaguchi, K. Miya, Y. S. Jong, E. Mathiowitz, N. K. Egilmez, R. B. Bankert.
2001
. Human inflammatory cells within the tumor microenvironment of lung tumor xenografts mediate tumor growth suppression in situ that depends on and is augmented by interleukin-12.
J. Immunother.
24
:
37
-45.
11
Hess, S. D., N. K. Egilmez, N. Bailey, T. M. Anderson, E. Mathiowitz, S. H. Bernstein, R. B. Bankert.
2003
. Human CD4+ T cells present within the microenvironment of human lung tumors are mobilized by the local and sustained release of IL-12 to kill tumors in situ by indirect effects of IFN-γ.
J. Immunol.
170
:
400
-412.
12
Broderick, L., S. J. Yokota, J. Reineke, E. Mathiowitz, C. C. Stewart, M. Barcos, R. J. Kelleher, Jr, R. B. Bankert.
2005
. Human CD4+ effector memory T cells persisting in the microenvironment of lung cancer xenografts are activated by local delivery of IL-12 to proliferate, produce IFN-γ, and eradicate tumor cells.
J. Immunol.
174
:
898
-906.
13
Lozupone, F., F. Luciani, M. Venditti, L. Rivoltini, S. Pupa, G. Parmiani, F. Belardelli, S. Fais.
2000
. Murine granulocytes control human tumor growth in SCID mice.
Int. J. Cancer
87
:
569
-573.
14
Santini, S. M., P. Rizza, M. A. Logozzi, P. Sestili, G. Gherardi, R. Lande, C. Lapenta, F. Belardelli, S. Fais.
1995
. The SCID mouse reaction to human peripheral blood mononuclear leukocyte engraftment: neutrophil recruitment induced expression of a wide spectrum of murine cytokines and mouse leukopoiesis, including thymic differentiation.
Transplantation
60
:
1306
-1314.
15
Yano, S., Y. Nishioka, K. Izumi, T. Tsuruo, T. Tanaka, M. Miyasaka, S. Sone.
1996
. Novel metastasis model of human lung cancer in SCID mice depleted of NK cells.
Int. J. Cancer
67
:
211
-217.
16
Tanaka, T., F. Kitamura, Y. Nagasaka, K. Kuida, H. Suwa, M. Miyasaka.
1993
. Selective long-term elimination of natural killer cells in vivo by an anti-interleukin 2 receptor β chain monoclonal antibody in mice.
J. Exp. Med.
178
:
1103
-1107.
17
Tanaka, T., M. Tsudo, H. Karasuyama, F. Kitamura, T. Kono, M. Hatakeyama, T. Taniguchi, M. Miyasaka.
1991
. A novel monoclonal antibody against murine IL-2 receptor β-chain: characterization of receptor expression in normal lymphoid cells and EL-4 cells.
J. Immunol.
147
:
2222
-2228.
18
Shultz, L. D., B. L. Lyons, L. M. Burzenski, B. Gott, X. Chen, S. Chaleff, M. Kotb, S. D. Gillies, M. King, J. Mangada, et al
2005
. Human lymphoid and myeloid cell development in NOD/LtSz-scid IL2Rγnull mice engrafted with mobilized human hemopoietic stem cells.
J. Immunol.
174
:
6477
-6489.
19
Serreze, D. V., J. W. Gaedeke, E. H. Leiter.
1993
. Hematopoietic stem-cell defects underlying abnormal macrophage development and maturation in NOD/Lt mice: defective regulation of cytokine receptors and protein kinase C.
Proc. Natl. Acad. Sci. USA
90
:
9625
-9629.
20
Serreze, D. V., H. R. Gaskins, E. H. Leiter.
1993
. Defects in the differentiation and function of antigen presenting cells in NOD/Lt mice.
J. Immunol.
150
:
2534
-2543.
21
Langmuir, P. B., M. M. Bridgett, A. L. Bothwell, I. N. Crispe.
1993
. Bone marrow abnormalities in the non-obese diabetic mouse.
Int. Immunol.
5
:
169
-177.
22
Traggiai, E., L. Chicha, L. Mazzucchelli, L. Bronz, J. C. Piffaretti, A. Lanzavecchia, M. G. Manz.
2004
. Development of a human adaptive immune system in cord blood cell-transplanted mice.
Science
304
:
104
-107.
23
Ito, M., H. Hiramatsu, K. Kobayashi, K. Suzue, M. Kawahata, K. Hioki, Y. Ueyama, Y. Koyanagi, K. Sugamura, K. Tsuji, et al
2002
. NOD/SCID/γcnull mouse: an excellent recipient mouse model for engraftment of human cells.
Blood
100
:
3175
-3182.
24
Yahata, T., K. Ando, Y. Nakamura, Y. Ueyama, K. Shimamura, N. Tamaoki, S. Kato, T. Hotta.
2002
. Functional human T lymphocyte development from cord blood CD34+ cells in non obese diabetic/Shi-scid, IL-2 receptor γ null mice.
J. Immunol.
169
:
204
-209.
25
Sugamura, K., H. Asao, M. Kondo, N. Tanaka, N. Ishii, K. Ohbo, M. Nakamura, T. Takeshita.
1996
. The interleukin-2 receptor γ chain: its role in the multiple cytokine receptor complexes and T cell development in XSCID.
Annu. Rev. Immunol.
14
:
179
-205.
26
Asao, H., C. Okuyama, S. Kumaki, N. Ishii, S. Tsuchiya, D. Foster, K. Sugamura.
2001
. Cutting edge: the common γ-chain is an indispensable subunit of the IL-21 receptor complex.
J. Immunol.
167
:
1
-5.
27
Cao, X., E. W. Shores, J. Hu-Li, M. R. Anver, B. L. Kelsall, S. M. Russell, J. Drago, M. Noguchi, A. Grinberg, E. T. Bloom, et al
1995
. Defective lymphoid development in mice lacking expression of the common cytokine receptor γ chain.
Immunity
2
:
223
-238.
28
Ishikawa, F., M. Yasukawa, B. Lyons, S. Yoshida, T. Miyamoto, G. Yoshimoto, T. Watanabe, K. Akashi, L. D. Shultz, M. Harada.
2005
. Development of functional human blood and immune systems in NOD/SCID/IL2 receptor γ chainnull mice.
Blood
106
:
1565
-1573.
29
Iwanuma, Y., F. A. Chen, N. K. Egilmez, H. Takita, R. B. Bankert.
1997
. Antitumor immune response of human peripheral blood lymphocytes coengrafted with tumor into severe combined immunodeficient mice.
Cancer Res.
57
:
2937
-2942.
30
Hess, S. D., N. K. Egilmez, J. Shiroko, R. B. Bankert.
2001
. Antitumor efficacy of a human interleukin-12 expression plasmid demonstrated in a human peripheral blood leukocyte/human lung tumor xenograft SCID mouse model.
Cancer Gene Ther.
8
:
371
-377.
31
Williams, S. S., T. Umemoto, H. Kida, E. A. Repasky, R. B. Bankert.
1992
. Engraftment of human peripheral blood leukocytes into severe combined immunodeficient mice results in the long term and dynamic production of human xenoreactive antibodies.
J. Immunol.
149
:
2830
-2836.
32
Chen, F. A., S. S. Williams, W. C. Fanslow, R. B. Bankert.
1995
. Human antibody response in human peripheral blood leukocyte/severe combined immunodeficient chimeric model is dependent on B and T cell costimulation via CD40/CD40 ligand.
J. Immunol.
155
:
2833
-2840.
33
Murphy, W. J., M. Bennett, M. R. Anver, M. Baseler, D. L. Longo.
1992
. Human-mouse lymphoid chimeras: host-vs.-graft and graft-vs.-host reactions.
Eur. J. Immunol.
22
:
1421
-1427.
34
Sasson, S. C., J. J. Zaunders, A. D. Kelleher.
2006
. The IL-7/IL-7 receptor axis: understanding its central role in T-cell homeostasis and the challenges facing its utilization as a novel therapy.
Curr. Drug Targets
7
:
1571
-1582.
35
Johnson, S. E., N. Shah, A. Panoskaltsis-Mortari, T. W. LeBien.
2005
. Murine and human IL-7 activate STAT5 and induce proliferation of normal human pro-B cells.
J. Immunol.
175
:
7325
-7331.
36
Barata, J. T., A. Silva, M. Abecasis, N. Carlesso, A. Cumano, A. A. Cardoso.
2006
. Molecular and functional evidence for activity of murine IL-7 on human lymphocytes.
Exp. Hematol.
34
:
1133
-1142.
37
Tary-Lehmann, M., A. Saxon.
1992
. Human mature T cells that are anergic in vivo prevail in SCID mice reconstituted with human peripheral blood.
J. Exp. Med.
175
:
503
-516.
38
King, M., T. Pearson, L. D. Shultz, J. Leif, R. Bottino, M. Trucco, M. A. Atkinson, C. Wasserfall, K. C. Herold, R. T. Woodland, et al
2008
. A new Hu-PBL model for the study of human islet alloreactivity based on NOD-scid mice bearing a targeted mutation in the IL-2 receptor γ chain gene.
Clin. Immunol.
126
:
303
-314.
39
Broderick, L., S. P. Brooks, H. Takita, A. N. Baer, J. M. Bernstein, R. B. Bankert.
2006
. IL-12 reverses anergy to T cell receptor triggering in human lung tumor-associated memory T cells.
Clin. Immunol.
118
:
159
-169.
40
Broderick, L., R. B. Bankert.
2006
. Membrane-associated TGF-β1 inhibits human memory T cell signaling in malignant and nonmalignant inflammatory microenvironments.
J. Immunol.
177
:
3082
-3088.
41
Nazareth, M. R., L. Broderick, M. R. Simpson-Abelson, R. J. Kelleher, Jr, S. J. Yokota, R. B. Bankert.
2007
. Characterization of human lung tumor-associated fibroblasts and their ability to modulate the activation of tumor-associated T cells.
J. Immunol.
178
:
5552
-5562.