Increased T Cell Glucose Uptake Reflects Acute Rejection in Lung Grafts



Although T cells are required for acute lung rejection, other graft–infiltrating cells such as neutrophils accumulate in allografts and are also high glucose utilizers. Positron emission tomography (PET) with the glucose probe [18F]fluorodeoxyglucose ([18F]FDG) has been employed to image solid organ acute rejection, but the sources of glucose utilization remain undefined. Using a mouse model of orthotopic lung transplantation, we analyzed glucose probe uptake in the grafts of syngeneic and allogeneic recipients with or without immunosuppression treatment. Pulmonary microPET scans demonstrated significantly higher [18F]FDG uptake in rejecting allografts when compared to transplanted lungs of either immunosuppressed or syngeneic recipients. [18F]FDG uptake was also markedly attenuated following T cell depletion therapy in lung recipients with ongoing acute rejection. Flow cytometric analysis using the fluorescent deoxyglucose analog 2-NBDG revealed that T cells, and in particular CD8+ T cells, were the largest glucose utilizers in acutely rejecting lung grafts followed by neutrophils and antigen-presenting cells. These data indicate that imaging modalities tailored toward assessing T cell metabolism may be useful in identifying acute rejection in lung recipients.


antigen-presenting cell




double costimulatory blockade




percent injected dose per cc of tissue






normalized mean fluorescence intensity


positron emission tomography


postoperative days


mean standard uptake value


three-dimensional volumes of interest.


Over the past 25 years, lung transplantation has become the treatment of choice for patients with end-stage lung disease. Although outcomes have improved over this time period, long-term survival remains disappointing. According to the latest International Society for Heart and Lung Transplantation (ISHLT) Registry report, the median survival in the most recent era (2000–2006) was 5.5 years [1]. Beyond the first year after transplantation, bronchiolitis obliterans syndrome (BOS), or chronic rejection, accounted for over 25% of deaths [1]. Acute rejection episodes are a primary risk factor for the development of BOS; even a single episode increases the risk of developing BOS [2]. Therefore, identifying acute rejection early is important to initiate treatment to reduce the risk of BOS.

Since mild and even moderate grade acute rejection can be clinically silent [3], surveillance transbronchial biopsies, the gold standard for diagnosing acute rejection, are performed at several centers. While these procedures are reasonably safe with experienced bronchoscopists, the number of biopsies required to reliably detect acute rejection can lead to increased risk of complications, including pneumothorax and bleeding [4]. Sampling error also ultimately limits the diagnostic sensitivity of this approach. Therefore, techniques that can improve on the detection of acute rejection would be highly useful for improving management and outcomes in these patients.

Positron emission tomography (PET) imaging with [18F]fluorodeoxyglucose ([18F]FDG) has been used to quantify lung inflammation [5-8] and may be a useful approach for quantifying acute rejection. Evidence in the literature suggests that [18F]FDG is taken up by activated immune cells, including T cells, which are the key mediators of acute lung transplant rejection [9]. T cells are known to take up glucose in response to activating stimuli to support the increased energy demands of the cell [10-12]. [18F]FDG uptake also increases with rejection in mouse models of acute rejection in lung, heart, kidney and liver transplantation [9, 13-15]. These data suggest that FDG-PET may be a useful approach for monitoring the efficacy of immunosuppressive therapy.

In this study, we characterized the time course of [18F]FDG uptake within the first 7 days after lung transplantation in an orthotopic left lung transplant mouse model. We hypothesized that in acutely rejecting lungs T cells are the major sinks for glucose uptake. Supporting our hypothesis, we demonstrated that recipients with ongoing acute lung allograft rejection have significant increases in [18F]FDG uptake driven primarily by the accumulation of T cells in the graft. In contrast, immunosuppression significantly reduced the overall sequestration of glucose tracer by the allogeneic lung graft T cell compartment, leading to decreased [18F]FDG uptake and thus allowing for clear PET-based discrimination of tolerant and rejecting lung grafts.

Materials and Methods

Animal groups and lung transplantation

All animal study procedures were approved by our institutional Animal Studies Committee. For the time course characterization experiments, C57BL/6 (B6) mice (male 6–10 weeks) received a left lung from either B6 (syngeneic lung graft) or Balb/c (allogeneic lung graft) mice and were imaged at Days 3 and 7 after transplantation. Separate cohorts of B6 mice receiving allogeneic grafts were either left untreated or treated with the following: double costimulatory blockade (DCB) immunoglobulins consisting of CD154 Ab (250 μg on postoperative day [POD] 0) and CTLA4-Ig (200 μg on POD 2) were administered intraperitoneally (i.p.). Cyclosporine (CsA) and methylprednisolone (MP) were given as single injections in the scruff behind the neck on the POD as indicated; CsA 5 mg/kg/day alone, (low dose treatment) CsA 5 mg/kg/day plus 0.8 mg/kg/day MP or (high dose treatment) 10 mg/kg/day CsA plus 1.6 mg/kg/day MP starting at the time of transplantation until sacrifice. These treatment cohorts were then imaged at Day 7 after transplant. A separate cohort of rejecting lung graft recipients received either 1 mg of Hamster isotype control or anti–thymocyte globulin treatment i.p. consisting of 0.5 mg GK1.5 (anti-CD4 Ab) and 0.5 mg YTS169.4 (anti-CD8a Ab) on POD 6 immediately after baseline imaging that day and then were imaged again the next day (POD 7). All lung transplantation procedures were performed as previously described [16, 17].

MicroPET and microCT imaging

Sixty–minute dynamic scans were acquired on an Inveon microPET/CT (Siemens, Siemens, Germany) microPET/CT or Focus 220 (Siemens/CTI) microPET scanner after injection of 124 ± 33 µCi of [18F]FDG. Computed tomography (Inveon) or transmission scans (Focus 220) were obtained for attenuation correction. Additional CT images were also obtained on those mice imaged on the Focus 220 microPET scanner. MicroPET images were reconstructed using two iterations of 3D ordered set expectation maximization and 18 of maximum a posteriori algorithm with beta value of 0.05, yielding 0.8 mm per pixel [18]. MicroCT images were acquired using 60 kVp, 350 ms per view with 220 views and reconstructed at 0.19 mm per pixel for attenuation correction and anatomic correlation.

Image analysis

Three-dimensional volumes of interest (VOIs) were placed over the lungs using the microCT and early blood flow microPET images as a guide with either ASIPro VM (Concorde Microsystems, Knoxville, TN) or Integrated Research Workflow 4.0 (Siemens) and lung time–activity curves extracted. We examined the percent injected dose per cc of tissue (%ID/cc) and mean standard uptake value (SUV) at 60 min postinjection in the lungs for quantifying the microPET [18F]FDG uptake. Because both of these measures yielded the same results in this particular model, we chose to quantify [18F]FDG uptake as the lung %ID/cc at 60 min posttracer injection.

Flow cytometric analysis and determination of 2-NBDG uptake in specific cell types

Lung recipients were fasted for 18 h and then sacrificed 45 min after intravenously injecting 750 μg of 2-deoxy-2-[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]-D-glucose (2-NBDG, Cayman Chemical, Ann Arbor, MI). Lung grafts were flushed to remove circulating pulmonary blood. Graft cell suspensions were prepared by collagenase digestion using previously described methods [16]. A T4 automated cell counter employing propidium iodide and cell debris discrimination software (Nexcelom Bioscience, Lawrence, MA) quantified the total number of live cells in intragraft single cell suspensions. The percent abundance of the specific cell types CD45+ (CD45.2+), CD4+ T (CD90.2+ CD4+), CD8+ T (CD90.2+ CD8+), B (CD19+), neutrophils (Gr1hi Ly6Ghi) and antigen–presenting cells (APCs) (CD11c+ CD11blo-hi MHCIIlo-hi) were quantified by FACS via staining with fluorochrome-labeled antibodies specific for CD45.2 (clone 104), CD19 (clone SJ25-C1), CD11b (clone M1/70), Gr1 (RBC-85), Ly-6G (clone 1A8), MHC-II (clone 28-5-16S), CD11c (clone HL3), CD90.2 (clone 30-H12), CD4 (clone RM4-5) and CD8 (clone 53-6.7) as well as their respective isotype controls (BD Pharmingen, San Jose, CA; eBioscience, San Diego, CA). The mean fluorescence intensity (MFI) determined with Flojo software (Version 7.6) at the excitation λ of 488 nm quantified 2-NBDG cellular uptake. The per cell 2-NBDG MFI for each specific cell type was normalized (normalized mean fluorescence intensity, NMFI) for the autofluorescence of the respective cell type by MFI analysis of control lung graft recipients that received 2-NBDG vehicle (saline) only. The NMFI multiplied by the total cell numbers per graft for each cell type (CD4+ and CD8+ T cells, B cells, neutrophils and APC) determined the contribution of each cell type to the total lymphoid and myeloid cell uptake of 2-NBDG in the graft. The per-cell NMFI of CD45+ cells multiplied by the number of CD45+ cells per graft determined the total 2-NBDG graft uptake from myeloid and lymphoid cells combined.


Lung graft tissue embedded in paraffin was stained with hematoxylin and eosin. Graft tolerance was evaluated by a blinded pathologist (D.C.) based on criteria established by the International Society of Heart and Lung Transplantation Working Lung Rejection Study Group of 2007 [19].

Statistical analysis

All data are represented as the mean ± SD. Statistical analyses were performed with SigmaStat 3.5 (Systat Software, Chicago, IL). Because each mouse had one transplanted allograft and one native lung, each lung was treated as an independent sample in all analyses. Two-way repeated measures analysis of variance tested for differences in [18F]FDG uptake at Days 3 and 7 posttransplant in allogeneic and syngeneic graft recipients (using time and graft type as factors) and in allogeneic graft recipients before and after anti-CD4/CD8 antibody treatment (using time and antibody treatment as factors). One-way ANOVA tested for differences in [18F]FDG uptake among DCB, CsA and untreated allogeneic graft recipients at POD 7. Two-way ANOVA tested for differences in 2-NBDG uptake among specific cell types within and among the different transplant condition groups, with cell type and treatment group as factors. The Tukey method for post hoc analysis identified significant interactions at p < 0.05.


Acute lung transplant rejection leads to increased [18F]FDG uptake

To assess the utility of FDG-PET for detecting acute lung transplant rejection, we compared microPET-measured [18F]FDG uptake in syngeneic lung graft recipients to allogeneic lung graft recipients in the absence of immunosuppression. Recipients receiving either syngeneic or allogeneic lung grafts had low levels of [18F]FDG uptake by POD 3. However, 7 days after engraftment, recipients with allogeneic grafts had visibly higher [18F]FDG uptake in the rejecting lungs than those who received syngeneic grafts, as evident by increased tracer activity in time–activity curves obtained from whole lung VOIs (Figure 1A and B). Consistent with our previous reports, POD 7 [16, 20] allogeneic graft histology showed clear evidence of acute lymphocytic vascular rejection as distinguished by prominent cuffing around pulmonary arterioles (Figure 1C). In contrast, Day 7 posttransplant syngeneic lung grafts had few histological signs of inflammation. Moreover, microPET–measured [18F]FDG uptake at Day 7 posttransplant was significantly increased compared to the [18F]FDG uptake measured in these same mice at Day 3 posttransplant and in the syngeneic graft recipients (Figure 2).

Figure 1.

Increased [18F]fluorodeoxyglucose ([18F]FDG) at Day 7 posttransplant in allogeneic lung allografts . Approximately 120 µCi of [18F]fluorodeoxyglucose ([18F]FDG) was injected in each mouse lung recipient at the beginning of a 60–min dynamic scan acquisition. (A) The last 5 min of the microPET images are displayed. The microCT images demonstrated variable degrees of consolidation among the different allogeneic recipients. However, all allogeneic recipients demonstrated increased [18F]FDG uptake the rejecting lung as illustrated. The syngeneic recipients uniformly had normally aerated lungs by microCT. (B) Volumes of interest placed over each lung determined the [18F]FDG time–activity curves in % injected dose per cc of lung (%ID/cc). (C) Hemotoxylin and eosin stained lung sections from syngeneic and allogeneic graft recipients at Day 7 posttransplant.

Figure 2.

[18F]FDG uptake increases from Days 3 to 7 posttransplant in allogeneic lung allografts . Syngeneic and allogeneic lung allograft recipients were imaged at both Days 3 and 7 posttransplant. The percent injected dose per cc of lung tissue (%ID/cc) at 60 min quantified [18F]FDG uptake in syngeneic (N = 3) and allogeneic (N = 5) graft recipients. Different grays for each square distinguish individual mice.

[18F]FDG uptake is significantly reduced in immunosuppressed lung allograft recipients

We next analyzed lung [18F]FDG uptake in allogeneic graft recipients immunosuppressed by several treatment regimens that predominantly target T lymphocytes to promote tolerance. To this end, allograft recipients were treated at the time of transplant with either a costimulatory blockade regimen that we have previously reported to induce lung acceptance in mice (DCB) [21], the calcineurin inhibitor cyclosporine (CsA), which is a commonly used immunosuppressant in the transplant clinic [22] as a single agent, or with CsA combined with MP at low or high doses. Representative images from mice treated with the low and high dose CsA + MP combination regimens are shown in Figure 3. Treatment with either DCB or CsA lowered lung allograft [18F]FDG uptake relative to transplanted lungs in recipients that did not receive immunosuppressive treatment (Figure 4). Notably, DCB treatment more effectively reduced [18F]FDG uptake than CsA treatment alone, in line with previous observations that CsA treatment alone can attenuate T cell activation but does not completely prevent acute rejection [23, 24]. Furthermore, high dose CsA in combination with MP was more effective in inducing tolerance compared with the low dose regimen, leading to more effective suppression of [18F]FDG uptake. The low dose CsA and MP treatment regimen reduced [18F]FDG similarly to low dose CsA alone, indicating that the addition of MP alone at these low doses did not significantly improve graft tolerance.

Figure 3.

Cyclosporine A (CsA) and methylprednisolone (MP) treatment reduces both acute rejection and [18F]FDG uptake levels . Top row: Lung allograft recipients treated with low dose CsA (5 mg/kg/day) and low dose MP (0.8 mg/kg/day) demonstrated partially reduced [18F]FDG uptake with variable amounts of infiltrate on the microCT images, correlating with acute rejection ISHLT Grade A3. Bottom row: Treatment with high dose CsA (10 mg/kg/day) and high dose MP (1.6 mg/kg/day) more effectively reduced [18F]FDG uptake than the low dose treatment regimen, correlating with minimal rejection severity ISHLT grade A1. The lungs more consistently appeared normal with minimal infiltrate by CT with the high dose regimen.

Figure 4.

[18F]FDG uptake varies in response to varying levels of immunosuppression in allogeneic lung allografts . Right panel: PET measured [18F]fluorodeoxyglucose ([18F]FDG_ uptake in native right (N) and allograft left (A) lungs, measured as % injected dose per cc of lung (%ID/cc) within the same mice in response to presence or absence of double co-stimulatory blockade (DCB) achieved by treatment with CD154 Ab (250 μg) and CTLA4-Ig (200 μg), cyclosporine A (CsA) and methylprednisolone (MP). Single + indicates low dose CsA (5 mg/kg/day) and MP (0.8 mg/kg/day). Double + (++) indicates high dose CsA (10 mg/kg/day) and MP (1.6 mg/kg/day). *p < 0.05 compared to untreated allograft recipients. †p < 0.05 comparing A to N within each treatment group. Of the mice left untreated, five of these data points are also represented in Figure 2; an additional two mice with imaging only at Day 7 posttransplant were included in this analysis. Left panel: Ratio of [18F]FDG %ID/cc in A:N lungs within each mouse. *p < 0.05 compared to untreated allograft recipients. †p < 0.05 compared to CsA alone. ‡p < 0.05 compared to low dose CsA + low dose MP.

Analysis of lymphoid and myeloid cell deoxyglucose uptake in lung allografts

As immunosuppression treatment lowered [18F]FDG uptake in lung allografts, we next sought to determine the differential contributions of graft-infiltrating lymphoid and myeloid cells to glucose uptake. Syngeneic and allogeneic recipients left untreated or treated with DCB were injected on postoperative Day 7 with the fluorescent glucose analog 2-NBDG 45 min prior to sacrifice and intragraft lymphoid and myeloid cells analyzed by flow cytometric analysis (Figure 5A and B). Uptake of 2-NBDG as assessed by NMFI demonstrated that T cells and APC on a per cell basis in untreated, rejecting allografts was significantly higher than that in the syngeneic grafts or DCB-treated allogeneic grafts. We next estimated the contribution of each graft–infiltrating cell population to 2-NBDG uptake per graft by multiplying the total number of graft-infiltrating cells (Figure 5C) by the respective NMFI (Figure 5D). T cells contributed the most to total graft 2-NBDG uptake, comprising over 75% of the total 2-NBDG uptake in rejecting allogeneic grafts (Figure 5D). Moreover, CD8+ T cells accounted for over half of the total 2-NBDG uptake for all graft-infiltrating cells in rejecting grafts (Figure 5D). Interestingly, while intragraft APCs and neutrophils in allogeneic lungs also had relatively high 2-NBDG uptake on a per cell basis when compared to the T cells, the contribution of these cell types to the total graft 2-NBDG uptake was significantly less than that of the T cells due to the lower numbers of these cells present in rejecting grafts relative to T cells. In contrast, B lymphocytes did not contribute significantly to 2-NBDG uptake under any of the studied transplant conditions. Not surprisingly, total 2-NBDG uptake for graft-infiltrating hematopoietic cells, as defined by CD45 expression, mirrored the FDG-PET results (Figure 5E vs. Figure 2). We further assessed the effect of low and high dose CsA and MP treatment on 2-NBDG uptake in graft-infiltrating hematopoietic cells. The combination of low dose CsA and low dose MP had little effect on any of the graft-infiltrating cells as the per cell 2-NBDG NMFI and numbers of graft-infiltrating cells per graft were similar to those found in untreated rejecting allogeneic grafts (Figure 6A). However, high dose CsA and high MP in combination reduced per cell glucose metabolism in CD8+ T cells to a greater extent than in CD4+ T cells, leading to reduced immune cell recruitment and overall reduced glucose metabolism in the graft (Figure 6B–D). Notably, PMN and APC glucose uptake on a per cell basis did not change in response to either of these treatment regimens.

Figure 5.

T cells are the predominant utilizers of 2-deoxyglucose in rejecting lung grafts . FACS analysis of 2-NBDG uptake of graft infiltrating cells in the transplanted lungs of syngeneic, allogeneic and DCB–treated allogeneic (allogeneic + DCB) recipients at Day 7 posttransplant. (A) Representative histograms (N = 6/group) for indicated graft–infiltrating cells from vehicle–treated (shaded) and 2-NBDG–treated (line) lung recipients. (B) 2-NBDG uptake MFI normalized (NMFI) to background autofluorescence measured in vehicle–treated recipients (N = 6/group). (C) Total number of indicated graft–infiltrating cells per lung graft (N = 6/group). (D) 2-NBDG uptake per lung graft (N = 6/group) for indicated graft–infiltrating cell type as determined by multiplying the total cell number shown in C by NMFI as shown in (B). (E) 2-NBDG uptake per lung graft for all graft–infiltrating (CD45+ cells) hematopoietic cells (N = 6/group). *p < 0.05 compared to syngeneic graft group. †p < 0.05 compared to allogeneic + DCB group. **p < 0.05 when comparing the group(s) under the darker bars to the groups under the lighter bars. Data represented as the mean + standard deviation.

Figure 6.

Cyclosporine A (CsA) and methylprednisolone (MP) treatment regulates T cell glucose uptake . FACS analysis of 2-NBDG uptake in graft-infiltrating cells from allogeneic graft recipients treated with either low dose (+) CsA (5 mg/kg/day) and MP (0.8 mg/kg/day) or high dose (++) CsA (10 mg/kg/day) and MP (1.6 mg/kg/day). *p < 0.05 comparing low versus high dose CsA + SM within each cell type for all panels. (A) Mean fluorescence intensity of 2-NBDG uptake normalized for background autofluorescence (NMFI) for each cell type. †p < 0.05 compared to CD4+ T cells and neutrophils (PMN) within high dose CsA + MP. ‡p < 0.05 compared to CD4+ T, PMN, antigen presenting cells (APC) within high dose CsA + MP and compared to all cell types within low dose CsA + MP. (B) Number of each cell type per lung graft. †p < 0.05 compared to B cells and PMN within high dose CsA + MP. Within the low dose CsA + MP group, all cell types were statistically different from the others except for CD4+ T compared to PMN and B cells compared to APC. (C) 2-NBDG uptake per lung graft (N ≥ 4/group) for indicated graft–infiltrating cell type as determined by multiplying the total cell number shown in (B) by NMFI shown in (A). No statistically significant differences between cell types were observed within the high dose CsA + MP group. All cell types were different from each other except CD4+ T compared to PMN and B cells compared to APC in the low dose CsA + MP group. (D) Total lung graft 2-NBDG uptake for all graft–infiltrating (CD45+ cells) hematopoietic cells. Data represented as the mean + standard deviation. §p < 0.05 compared to B cells and PMN.

Anti–thymocyte immunoglobulin treatment of an acute lung rejection episode reduces [18F]FDG uptake

T cell depletion with anti–thymocyte immunoglobulin treatment is a clinical approach used to treat refractory acute lung rejection episodes [25]. Therefore, we further asked if we could monitor changes in lung allograft rejection via [18F]FDG uptake with this treatment strategy. Mice receiving allogeneic lung grafts were imaged at Days 6 and 7 posttransplant, before and after treatment with either anti–thymocyte antibodies or isotype control antibodies. MicroPET imaging with [18F]FDG demonstrated stably increased uptake in recipients treated with the isotype control antibody (Figure 7A). In contrast, recipients treated with the anti–thymocyte antibodies demonstrated decreased [18F]FDG uptake on Day 7 imaging, correlating with histological evidence of decreased intragraft inflammation (Figure 7B).

Figure 7.

FDG-PET detects effective immunosuppressive treatment of acute rejection . Mice imaged on Day 6 posttransplant were treated with a single dose of T cell depleting (N = 5) or isotype control antibodies (N = 5) and then imaged again the day after treatment (POD 7). Lung sections were then obtained to assess for acute rejection. (A) The percent injected dose per cc lung tissue (%ID/cc) at 60 min quantified [18F]FDG uptake. (B) Lung sections from mice treated with the isotype control antibody versus T cell depleting antibody. White circles: native lung. Gray circles: allogeneic lung grafts. *p < 0.05 compared to native lung. **p < 0.05 compared to POD 6.


Our imaging data in a mouse model of orthotopic lung transplant suggest that FDG-PET may be useful non-invasive approach to detect acute rejection. When compared to mice that received syngeneic lung grafts, mice with acutely rejecting allogeneic lung grafts had markedly elevated [18F]FDG uptake that correlated with histological evidence of rejection. The syngeneic lung grafts also did not exhibit increased [18F]FDG uptake but instead had utilization that was similar to the native lungs. Our study thus agrees with published studies in other lung and solid organ transplant models that demonstrate increased [18F]FDG uptake with acute rejection [9, 13-15].

To determine if FDG-PET could be used to monitor tolerance in lung recipients, we immunosuppressed allograft recipients. Notably, DCB treatment nearly eliminated the [18F]FDG uptake that normally would have occurred as a result of rejection, suggesting that T cells in rejecting grafts play a significant role in driving glucose utilization under these conditions. Although T lymphocytes are necessary for acute lung rejection [16], these grafts also accumulate large numbers of myeloid cells [26], which are likewise potential sources of glucose sequestration. To better characterize [18F]FDG allograft signals we employed 2-NBDG [27], a fluorescent 2-deoxyglucose probe that has been used as a sensitive surrogate measure of glucose uptake in T, B [28] and myeloid cells [29] in vivo. We adapted the use of this probe in a flow cytometric–based assay to forward the study of differential glucose utilization by several graft–infiltrating myeloid and lymphoid cell populations. We observed that patterns of total 2-NBDG uptake by graft–infiltrating cells was comparable to [18F]FDG uptake by microPET under these same conditions.

Both CD4+ and CD8+ T cells are dependent on co-stimulation signals that drive lymphocyte proliferation leading to solid organ rejection [30]. Notably, the CD28 co-stimulation pathway is a critical regulator of T cell proliferation leading to the expression of glucose transporters such as Glut1 and the subsequent upregulation of glucose uptake [31]. Under glucose limiting conditions, proliferating T cells rapidly undergo apoptosis [32], which is in line with previous observations that CD28 costimulation blockade promotes the activation–induced T cell apoptosis known to enable graft tolerance. [33, 34]. To study the impact of this approach on glucose uptake in T cells specifically under tolerant conditions, we used a DCB immunosuppressive regimen, which is dependent on CD28 blockade to promote lung allograft acceptance [21]. DCB treatment not only significantly lowered glucose uptake in individual T cells but also attenuated levels of intragraft T cell abundance to levels comparable to that observed in syngeneic grafts. Additionally, we could detect changes in intragraft T cell glucose uptake by modulating the amount of immunosuppression administered to lung recipients. Both allograft [18F] FDG and T cell 2-NBDG uptake was significantly lower in high dose CsA and MP treated recipients when compared to low dose treated recipients. Finally, because the number of activated T cells were still more numerous than the other cell types present in rejecting lungs, the combination of both higher per cell glucose uptake and increased accumulation suggests that T cells are the primary source of [18F]FDG uptake during an acute rejection episode, regardless of the severity of inflammation.

Our detailed FACS analysis of 2-NBDG uptake in rejecting lungs also revealed novel insights about the glucose uptake responses of non–T cell hematopoietic cell types during acute rejection–associated inflammation. While other parenchymal cell types within the lungs are known to take up glucose under various inflammatory conditions [35, 36], the collagenase digestion method used to release resident hematopoietic cells destroys parenchymal tissue; therefore, analysis of these cell types was not possible with this method. We clearly observed 2-NBDG uptake in neutrophils and APCs on a per cell basis in addition to that observed in the T cells. Surprisingly, neutrophils took up similar amounts of 2-NBDG per cell irrespective of whether they were infiltrating syngeneic, allogeneic or DCB–treated allografts and regardless of the presence or absence of steroids. We have shown that glucose uptake increases in activated neutrophils entering extravascular spaces in human lungs [37] and that increase glucose uptake can be detected by PET even in the initial steps of activation, before neutrophils enter the airspace [6]. As neutrophils must be activated to enter extravascular spaces [38], the observations from the current study could be the result of constitutive surveillance mechanisms that drive small numbers of neutrophils into syngeneic and tolerant allografts, in line with our previously published data with intravital 2-photon microscopy demonstrating low levels of constitutive neutrophil trafficking into non–inflamed lungs [39]. Additionally, our results clearly demonstrate that neutrophil glucose uptake on a per cell basis is resistant to combination CsA and MP treatment. In particular, corticosteroids may have diametric effects on neutrophil metabolism since these agents are known to attenuate reactive oxygen intermediate production and degranulation but also halt apoptosis [40], which would suggest some overall demand to maintain glucose utilization. Although glucose metabolism by individual neutrophils was unaffected by immunosuppression levels, high dose CsA and MP had a dramatic impact on preventing lung allograft neutrophil infiltration. This may be the result of two complementary effects on neutrophil recruitment. CsA prevents naïve CD4+ T cells from differentiating into Th17 cells [41], a T cell subset found in acutely rejecting lungs [21] that promotes neutrophil accumulation by stimulating granulocyte production and mobilization [42], while corticosteroids work further downstream, acting as transcriptional regulators that inhibit the expression of chemokines and adhesion molecules that promote extravascular accumulation [43, 44]. In contrast, we observed that B cells took up little 2-NBDG. The reasons for this are unclear as B cells express several glucose transporters [45]. However, as opposed to T cell and neutrophil activation, which can be initiated in the lung [20, 46, 47], naïve B cell activation may mostly be restricted to draining secondary lymphoid organs and thus contributes little to intragraft glucose utilization during acute rejection.

We also evaluated the effect of administering low dose CsA at the time of transplant on [18F]FDG uptake in allogeneic lung grafts as well as low and high doses of CsA in combination with low and high dose MP treatment. The absolute [18F]FDG uptake decreased relative to untreated recipients but not as markedly as that seen with DCB treatment. Interestingly, the addition of low dose MP to CsA therapy did not improve the level of [18F]FDG uptake suppression; however, high dose CsA in combination with high dose MP more effectively reduced [18F]FDG uptake. Blunted antigen presentation mediated by primarily MP may partially explain the superior effectiveness of high dose CsA and MP treatment. Unlike CsA, high dose corticosteroids have been reported to prevent the upregulation of MHC molecules and costimulatory signals on dendritic cells known to drive IL-2 expression and proliferation of naïve T lymphocytes [48]. Accordingly, a recent report suggests that CsA treatment alone is not sufficient to prevent lung graft rejection in mice [24].

To further test the ability of FDG-PET as such a biomarker, we treated allogeneic lung allograft recipients with anti–thymocyte globulin antibodies at Day 6, a time point at which acute rejection is already established, and obtained baseline pretreatment and immediate (approximately 24 h) posttreatment FDG-PET imaging. We observed a consistent reduction in the FDG-PET signal in animals treated with active antibody compared to those treated with the isotype control that again correlated with histological evidence of reduced T cell activation. These results support the potential for using FDG-PET to monitor responses to immunosuppressive interventions for acute rejection. Such an application for FDG-PET will most likely be most useful in patients who do not have other evidence of infection. Given our data demonstrating that neutrophil glucose utilization is not affected by immunosuppression, infectious processes would need to be excluded to improve the confidence of interpreting a lack of change in [18F]FDG uptake as a failure of immunosuppression therapy as opposed to superimposed infection. Thus, interpretation of the FDG-PET study for clinical applications should still occur within the context of other clinical information, including microbiology, biopsy and BAL results. Despite this limitation, FDG-PET may still provide useful information about the extent of rejection (e.g. confirming other areas potentially involved in the acute rejection process when biopsy and BAL results confirm the presence of acute rejection without superimposed infection). FDG-PET may even be potentially useful in directing which segments to interrogate by bronchoscopy, thus improving the diagnostic yield of bronchoscopic procedures. Finally, the development of novel tracers that can target different immune cells more specifically could be very helpful in differentiating infection from rejection.

While such clinical applications for FDG-PET are promising, the potential sensitivity of FDG-PET for detecting acute rejection must be determined to establish how FDG-PET might be used in the clinical management of lung transplant recipients. The only human study evaluating FDG-PET for imaging lung transplant recipients quantified [18F]FDG uptake using the Patlak graphical analysis, a model–free approach for quantifying kinetic PET data [49-51]. In that study, the rate of [18F]FDG uptake in the one subject with evidence of A2 rejection was at the upper range of values reported in healthy volunteers to date [5, 8, 52]. In our experience, such kinetic analyses in human studies appear to be more sensitive for quantifying differences in low levels of lung inflammation [53], unlike the %ID/cc used in this study or related SUV that is used clinically. However, dynamic imaging is used more frequently for research applications and rarely used for clinical FDG-PET scans. Thus, evaluating whether more clinic–friendly image acquisition protocols can provide the same information would also be helpful in establishing the potential utility of this technique for clinical practice.

In conclusion, we have demonstrated that FDG-PET may be useful biomarker of acute lung transplant rejection in a mouse model of orthotopic lung transplant. Our data support further assessing the utility of this approach for quantifying acute rejection in human lung transplant recipients.


The Mallinckrodt Institute of Radiology provided departmental funding for the imaging studies. DLC's effort was funded by NIH K08 EB006702, NIH R01 HL116389 and the Doris Duke Charitable Foundation Clinical Investigator Award. AEG's effort was supported by NIH R01 HL094601-01, R01 HL113931-01, R01 HL113436-01A1 and the Barnes Jewish Hospital Foundation. The authors thank Nicole Fettig, Amanda Roth, Margaret Morris, Lori Strong and Ann Stroncek of the MicroPET Facility for performing the imaging acquisitions and the Cyclotron Facility for radiopharmaceutical production.


The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.