Sustained Anti-CD4/CD8 Treatment Blocks Inflammatory Activation and Intimal Thickening in Mouse Heart Allografts

Chronic rejection manifested by arterial occlusion and parenchymal fibrosis limits the long-term survival of transplanted hearts. The classic features of transplant arteriosclerosis, which develops faster than naturally occurring atherosclerosis, include diffuse, concentric, and hypercellular arterial lesions that occur only within the donor heart.¹² These vascular lesions are common features of most solid-organ transplants undergoing chronic rejection. Review of clinical studies suggests that acute cellular rejection and the number of HLA antigen mismatches may be risk factors for development of transplant arteriosclerosis or vasculopathy.³ Other antigen-dependent and independent factors have also been implicated in the pathogenesis of vascular thickening.⁴

Tremendous insight into antigen-dependent factors and transplant arteriosclerosis has been gained from animal models with long-term allograft survival and vascular lesions that mimic those seen in human allografts.⁵⁶ Rodent models are inexpensive, pure genetic strains are available, and transplant arteriosclerosis develops rapidly in them. Donor vessels in allografts develop dramatically more intimal thickening than host or isografts vessels. This supports the hypothesis that an alloimmune response initiates and propagates intimal thickening in the long-term surviving grafts. However, the precise cellular and molecular pathways that culminate in the vessel changes of chronic rejection, and their relations to acute rejection, are still under investigation.

Mice are ideally suited for study of the immunologic pathways of chronic rejection because they are well-characterized genetically and immunologically and the repertoire of genetically manipulated mice is ever growing.⁵ Recent studies demonstrate that short courses of monoclonal antibodies in the early posttransplantation period can produce prolonged graft survival and development of intimal thickening in a number of strain combinations.⁷⁸⁹¹⁰ To establish a model to explore specific inflammatory pathways associated with lesion development we studied mouse transplants with major histocompatibility complex mismatches by placing CBA donor hearts into C57 recipients. These recipients were selected specifically to provide a basis for future studies with genetically modified recipients typically generated on a 129/C57 background. In this study, long-term surviving grafts from recipients treated with a short course of anti-CD4 were compared with grafts from recipients treated with the more aggressive regimen of anti-CD4 and anti-CD8, which prevents thickening.

Methods

Mouse Heart Transplantation and Treatment

Sixty-six vascularized, heterotopic abdominal cardiac transplantations were performed in mice.¹¹¹² CBA/CaJ (H-2^k) donors and C57BL/6J (H-2^b) recipients (The Jackson Laboratory, Bar Harbor, Maine) were used in the allograft combination (n=49). The C57BL/6J (H-2^b) strain was used for isografts (n=17). Either a short course of treatment with CD4 antibody (GK1.5; 0.5 mg intraperitoneally, days 1 to 4, n=33 allografts) or a sustained course with CD4 and CD8 antibodies (GK1.5 and 2.43, respectively; 0.5 mg intraperitoneally, days 1 to 4 and then weekly until harvest, n=13 allografts) were administered in an effort to obtain grafts that survived longer. Three of the 49 allografts were left untreated. We initially administered the short course of anti-CD4 because long-term graft survival has been reported with this program in other strain combinations.¹³ However, after learning in preliminary studies of the CBA-to-C57 combination that anti-CD4 alone failed to uniformly prevent acute rejection, we decided to administer the sustained, combined anti-CD4/CD8 treatment. Five of the 17 isografts also received sustained anti-CD4/CD8 treatment. All grafts were evaluated daily for ventricular function by measuring the force of palpable heart.¹¹ Grafts were harvested when the palpation score was 1 (on a scale of 0 to 4) or when the animal had survived for more than 95 days (the definition of a long-term survivor). Four of the 12 untreated isografts were harvested 365 days after transplantation to evaluate the possibility of late intimal thickening.

Purification of CD4 and CD8 Antibodies

Monoclonal antibodies against CD4 (GK1.5,¹⁴ ) and CD8 (2.43,¹⁵ ) were purified from the Pristane-induced (2,6,10,14-tetramethylpent-adecane, Pristane, Sigma, St. Louis, Mo) ascites of nude mice. Immunoglobulin was purified on UltraLink immobilized protein G columns (Pierce Chemicals, Rockford, Ill) according to the manufacturer’s protocol. Depletion of CD4 and CD8 cells was demonstrated by flow-cytometry analysis of splenic tissue harvested from mice 7 days after a 4-day-treatment was ceased, or at the time of transplant harvest (100 days). The FACS analysis showed greater depletion of CD4⁺ cells (3% of total leukocytes) and CD8⁺ cells (5%) 7 days after cessation of treatment compared with naive C57 mice (23% and 27%, respectively). At 100 days the depletion continued (6% of CD4⁺ cells and 8% of CD8⁺ cells) in the recipients who received long-term treatment.

Histologic Study and Morphometric Analysis

Morphometric analysis was performed on 37 long-term surviving grafts (>95 days). After perfusion with phosphate-buffered saline, sagittal heart sections were fixed in methyl Carnoy’s solution and embedded in paraffin. The histologic features of the transplanted hearts were then evaluated with hematoxylin and eosin staining. Sections (4 μm) were stained with Verhoeff’s elastin to visualize the internal elastic lamina and to aid in the identification of intimal lesions in intramural arteries. The severity of disease (percentage of luminal occlusion and morphometric score), as well as the frequency of disease (percentage of diseased vessels), was analyzed in two sections from each graft. Microscopic images of each elastin-stained vessel were captured (total number, 371; mean, 11.6 vessels/graft) and the percentage of luminal occlusion was tabulated by tracing the internal elastic lamina and the lumen with the ScionImage 1.59 software (National Institutes of Health, Bethesda, Md).¹⁶ The morphometric scores were 0 (normal), 1 (<20% luminal occlusion, <50% circumferential vessel involvement), 2 (<20% occlusion, >50% circumferential vessel involvement), 3 (20% to 50% occlusion, >50% circumferential vessel involvement), 4 (50% to 80% occlusion, 100% circumferential vessel involvement), and 5 (>80% occlusion, 100% circumferential vessel involvement).¹⁷ The morphometric score and the percentage of luminal occlusion were both analyzed by two independent examiners (r=.96 and r=.92, respectively, P<.0001 for both), and the mean value for each individual graft was tabulated. The frequency of the disease in the graft was defined as percentage of diseased vessels (>0% luminal occlusion). The mean±SD value for all grafts in each transplant subgroup is reported.

Gene Expression Measured by RT-PCR

Total cellular RNA was extracted from snap-frozen ventricular sections of long-term grafts with RNAzol B (Tel-Test, Inc, Friendswood, Tex). The quality of the RNA was confirmed on formaldehyde gels. cDNA synthesis was performed according to the manufacturer’s protocol (cDNA kit, Gibco BRL Life Technologies, Grand Island, NY). We used a published RT-PCR technique¹⁸ to measure relative differences in transcript levels after normalization against levels of the reference gene G3PDH. Primer sequences, which crossed exon-intron borders whenever possible, were selected with the MacVector 4.1.1. software (Eastman Chemical Co, New Haven, Conn) and synthesized on the Oligo 1000 DNA Synthesizer (Beckman Instruments, Fullerton, Calif). For each primer combination, conditions were optimized to generate a single specific band. Amplification of the predicted fragment was confirmed using restriction mapping. The GeneAmp 9600 system (Perkin Elmer, Foster City, Calif) was used to establish logarithmic ranges of PCR amplification as a function of cycle number and cDNA dilution, and the hot-start technique was used to increase specificity.¹⁹ Reaction conditions included 1.25 μL cDNA, 1 μmol/L each of 5′ and 3′ primer, 10 mmol/L Tris hydrochloric acid, 50 mmol/L of potassium chloride, 1.5 mmol of magnesium chloride, 0.001% (wt/vol) gelatin, 800 μmol/L dNTP (200 μmol/L of each), and 0.625 U AmpliTaq DNA polymerase (Taq DNA polymerase, Perkin Elmer, Norwalk, Conn) in a total volume of 25 μL. ³²P-dCTP (150,000 cpm) was included for quantitative PCR studies. The thermal cycling parameters included denaturation at 94°C for 15 seconds, annealing between 50°C and 60°C for 20 seconds, and extension at 72°C for 60 seconds (with a final extension of 7 minutes at the end of all cycles). The specificity of the MCP-1 primers was increased by using the touchdown PCR technique,¹⁹ which included a gradual decrease in the annealing temperature from 70°C to 65°C over 10 cycles and using an annealing temperature of 56°C for an additional 18 cycles thereafter. Accession numbers, primer sequences, annealing temperatures, and number of cycles were as follows.

PCR products (10 μL) were analyzed on 1% to 2% agarose gels, and incorporation of ³²P-dCTP into PCR product bands was measured from dried gels on a PhosphorImager (Molecular Dynamics, Sunnyvale, Calif), as previously described.¹⁸ PCR amplification with the G3PDH housekeeping gene was performed to assess variations in cDNA or total RNA loading between samples. The mean obtained from at least three analyses was used to normalize other transcript level measurements obtained from the same sample. Corrected values were derived by dividing the measured ³²P value for the transcript of interest by the mean G3PDH value for the sample. Relative transcript levels were then determined from cDNA panels that included a negative control (in which water was used for the PCR instead of cDNA). Because of limited amounts of RNA and therefore cDNA, PCR was performed on two to eight samples in each subgroup (numbers are shown in Figures 1 through 6) to identify relative differences; each analysis was performed at least three times. Because RT-PCR does not provide information about protein levels or localization, we selected inflammatory factors that had been identified previously by immunostaining in mouse grafts.^8132324

Statistical Analysis

Results are given as the mean±SD per subgroup, which was derived from the mean per graft. The data were subjected to MANOVA without replication (StatView 4.1, Abacus Concepts, Inc, Berkeley, Calif). If the results of the MANOVA were significant, individual comparisons were made with the Student’s t test and the level of significance was corrected using the Bonferroni method. To evaluate potential correlations between different parameters, we subjected the correlation coefficients to Fisher’s r to z transformation.

Results

Survival

Untreated C57 mice that received CBA hearts rejected the allografts within 8.7±0.7 days (no long-term survivors, n=3) (Fig 1). Sustained, combined anti-CD4/CD8 treatment significantly prolonged mean survival times of CBA to C57 allografts to 101±20 days (85% experienced long-term survival, n=11/13). This survival time is similar to that obtained in the C57 syngeneic combination (n=17), in which all grafts survived longer than 95 days. In contrast, a short course of anti-CD4 alone produced a CBA to C57 allograft survival time of only 48±30 days (27% had long-term survival, n=9/33, P<.0001 compared with combined treatment group or isografts).

Morphologic Analysis

Histological analysis of cardiac allograft sections from long-term survivors (>95 days) in the group treated with a short course of anti-CD4 showed patchy mononuclear cell infiltration and fibrosis in the parenchyma. The donor vasculature showed perivascular infiltration, occasional disruption of the internal elastic lamina, and intimal thickening (Fig 2, top left). In contrast, the combined anti-CD4/CD8–treated group and isografts (Fig 2, top center and right) had minimal cell infiltration within the parenchyma and perivascular space; intimal lesions were infrequent and generally minor. A detailed evaluation of lesion severity in all elastin-stained vessels showed a heterogeneity in the degree of intimal thickening within a single graft which was independent of the treatment group. However, the allografts from mice that received the short course of anti-CD4 exhibited a distribution toward more severe lesions (Fig 2, bottom). The converse was true in the isografts and the allografts that received the combined anti-CD4/CD8 treatment.

A precise calculation of the percentage of luminal occlusion in each vessel (n=371; Fig 3A) showed significantly more severe disease in allografts from mice that received the short course of anti-CD4 (51.6% mean occlusion) than in those from mice that received sustained anti-CD4/CD8 (8.3% mean occlusion) or isografts (6.9% mean occlusion; P<.0001). Lesion severity measured with morphometric score (scale 0 to 5; Fig 3B) was also significantly higher in the anti-CD4–treated group (3.1±0.8) than in the sustained anti-CD4/CD8–treated group (0.5±0.6) or the isograft group (0.5±0.5; P<.0001). These qualitative measurements of lesion severity (morphometric scores) correlated highly with the quantitative measurements (luminal occlusion assessed with the aid of a computer) ( r=.98; P<.0001).

Lesion frequency, measured as a percentage of diseased vessels (Fig 3C), followed a similar pattern. In the short-course group some thickening was noted in 94.1% of the arteries, whereas only 21.1% of the arteries in the sustained treatment group and 15% of the arteries in isografts showed any disease. Thus, the group that received sustained anti-CD4/CD8 treatment through the duration of the graft showed significantly less frequent and less severe transplant arteriosclerosis than did the group receiving the short course of anti-CD4. Lesion development in the group that received sustained anti-CD4/CD8 therapy was more similar to that in the isograft group.

In the five isografts treated with the sustained combination of anti-CD4/CD8 there was no significant difference in the severity of disease measured with intimal occlusion (Fig 3A inset, mean of 9.0%) or morphometric score (Fig 3B inset, mean score of 0.6). In the four untreated isografts harvested 365 days after transplantation, the percentage of luminal occlusion and morphometric score (Fig 3A and B, insets, 13% and 0.7, respectively) were not significantly different from those for the eight untreated isografts harvested after 95 days (2.5% and 0.2, respectively).

Expression of Adhesion Molecules and Cytokines

We measured gene transcript levels with RT-PCR to compare inflammatory cell infiltration and activation among subgroups of grafts that survived longer. The proximal aspects of the inflammatory cascade were evaluated by measuring cytoadhesion molecules, which are thought to initiate the inflammatory response.²⁵ Fig 4 shows that the mean transcript levels for ICAM-1 and its ligand, LFA-1, were sixfold (P=.006) and eightfold (P=.002) higher in the group treated with a short course of anti-CD4 than in the group treated with a sustained course of anti-CD4/CD8 or the isografts. Relative transcript levels were significantly higher for P-selectin (24-fold, P=.0009) but not for E-selectin (2-fold, P=NS) in the anti-CD4–treated group. Fig 5 shows the cytokines that we evaluated. These are believed to amplify the inflammatory response, and they have been identified previously in grafts undergoing acute rejection. Compared with the sustained anti-CD4/CD8–treated group and isografts, transcript levels increased significantly in the anti-CD4–treated group for IL-4 and IFN-γ (by sevenfold each, P<.001 and P<.003, respectively) but not for IL-2 (twofold, P=NS). Fig 6 shows the transcript levels for factors typically produced by activated macrophages. Again, levels were highest in the group treated with the short anti-CD4 course for AIF-1 (9.1-fold, P=.005), MCP-1 (4.7-fold, P=.0007), and iNOS (29-fold, P<.0001).

Correlation Between Histological Findings and Gene Expression

To evaluate the association between intimal thickening in the vessels of transplanted hearts and transcript levels of mediators of inflammation, we plotted the percentage of luminal occlusion against the transcript levels (data not shown). A significant correlation was found between the percentage of intimal thickening and ICAM-1 (r=.86, P<.0001), LFA-1 (r=.93, P<.0001), P-selectin (r=.89, P<.0001), E-selectin (r=.88, P<.01), IL-2 (r=.86, P<.01), MCP-1 (r=.93, P<.0001), and AIF-1 (r=.92, P<.0001) transcript levels. Interestingly iNOS and IFN-γ had lower association with luminal occlusion (0.761, P<.046, and 0.732, P<.062). There was also a high association between the transcript levels for IL-2 and AIF-1 and those for the adhesion molecules (r=.92–0.99, P<.0001), and among those for the adhesion molecules themselves (r=.94–0.98, P<.0001).

Discussion

In this study, we show for the first time, to our knowledge, direct correlation between expression of a number of inflammatory factors (cytoadhesion molecules, cytokines, and macrophage activation products) in a large series of transplantations in which vessels were analyzed using quantitative techniques. Extensive vascular thickening develops after a short course of anti-CD4 in major histocompatibility complex class I and II disparate mouse hearts placed into C57 recipients that survived long term. The short treatment prolongs graft survival and attenuates acute rejection, but still results in mononuclear infiltration and vascular lesion formation in long-term surviving grafts. Most of these inflammatory activators were selected for study because they had been previously identified in acutely rejecting grafts.²⁶²⁷ The presence of the same factors in grafts undergoing chronic rejection supports the concept that the vascular lesions of chronic rejection may develop as part of a continuum of acute rejection. However, this hypothesis does not rule out other contributing factors.

The findings in this C57-based model that transient or short-term therapies can be used to control acute rejection and produce grafts with vascular thickening supplements findings from other reports involving a variety of mouse strain combinations.^78102829 We also demonstrate that lesion development is inhibited by continuous blockade with anti-CD4 combined with anti-CD8. Hence we show that long-term graft survival does not predict the development of chronic vascular lesions in mouse models. In addition, we confirm that a T-cell driven immune response is required for vascular thickening as suggested by recent studies involving various types of rodent transplant models. Shi et al³⁰ showed that carotid artery loops placed into allogeneic recipient mice deficient in CD4⁺ cells developed significantly smaller lesions than did loops placed into control recipients with normal levels of CD4⁺ cells. In our heart transplant study, it is not clear whether vascular thickening was impaired by the protracted nature of the blockade or the addition of anti-CD8, or both. We suspect that inhibition of CD8 in our mouse heart allografts may not have been critical to the prevention of thickening, because lesion development in the carotid artery loops was not altered in recipient mice deficient in CD8⁺ cells.³⁰ Other strategies using ongoing combination treatment with anti–ICAM-1 and LFA-1³¹ or sustained treatment with gallium nitrate²⁸ caused a significant decrease in vascular lesions in mouse allografts. Taken together, the attenuation of vascular thickening in our studies probably arises from the ongoing disruption of T-cell mediated responses and not just the addition of anti-CD8. Selective depletion of CD4 or CD8 individually with monoclonal antibodies or gene targeting will clarify this issue.

A growing number of alloantigen-independent factors have been identified as important contributors to chronic rejection. For example, Tullius et al³² found that after 24 weeks, kidney isografts or single ischemic kidneys develop arteriosclerosis, glomerulosclerosis, interstitial fibrosis, and tubular atrophy in association with late mononuclear cell infiltration and cytokine expression. To determine whether murine cardiac allografts would undergo a similar late response to the initial surgical injury, we compared isografts harvested at 13 weeks with those harvested at 52 weeks after transplantation. At both times, isografts showed only occasional and mild vessel thickening without a significant difference in mean percentage luminal thickening values. This suggests that in heterotopic heart transplant models, early effects of ischemia/reperfusion are not further amplified over time as they might be in kidney models in which the diminished renal mass or number of nephrons is associated with hyperfiltration and fibrosis.³²³³

The concept that intimal thickening after transplantation arises, in part, from the activation and amplification of T cell and macrophage factors is supported by our RT-PCR data. We originally anticipated that only a subset of inflammatory factors would be upregulated in long-term grafts. This outcome would imply that only certain inflammatory factors were induced in chronic rejection. Instead we found that transcript levels for the selected cytoadhesion molecules (ICAM-1, LFA-1, and P-selectin), T cell–derived cytokines (IL-4, and IFN-γ), and factors expressed by activated macrophages (MCP-1, iNOS, and AIF-1) were all significantly higher in the grafts that developed significant arteriosclerosis after treatment with the short course of anti-CD4. IL-2 and E-selectin were the only two factors whose transcript levels were not significantly higher in the short-course group. For these two, the levels were still higher than in the short course group but the degree was smaller. Hence, larger numbers might be required to achieve statistically significant differences. In this study, there was a linear correlation among all inflammatory factor transcript levels and the severity of intimal thickening. Similar patterns of coordinate upregulation of such inflammatory factors have also been observed in other rodent models of chronic rejection, with the process of acute rejection, and with other inflammatory diseases such as arthritis,³⁴ encephalitis,³⁵³⁶ and ischemic acute tubular necrosis.³⁷ Taken together, this suggests that there may be a common multicellular inflammatory response to injury which produces late fibrotic manifestations (vascular thickening being one) that vary depending on the organ and duration of the injury.

This C57-based mouse model of transplant vasculopathy shares many of the previously described features of chronic rejection found in other animal models—diffuse, concentric, vascular thickening develops in a spectrum of vessels in allografts compared with isografts.⁵ Intimal thickening has been described to varying extents in rodent models in association with assorted inflammatory mediators and growth factors.^2203839 Finally, our studies using sustained antibody treatment suggest that vascular thickening appears to be mediated in part by a T cell–driven response. Here we provide quantitative ways of assessing both vascular thickening and inflammatory activation. The next and more arduous task will be to determine which inflammatory factors play a causal role in the development of vessel thickening after transplantation. Given that chronic inhibition of such factors would be required, one approach is to use knockout mice in transplant studies. Such studies have been initiated in acute heart transplant models evaluating targeted deficiencies of ICAM-1,⁴⁰ IL-2 or both IL-2 and IL-4,⁴¹ or IFN-γ⁴² where the endpoint is graft survival but not chronic rejection. The C57-based mouse model described herein can be adapted to study selected immune deficiencies of targeted factors using knockout mice as recipients or donors. If a gene is deleted and the arterial walls of the altered mouse do not thicken, then we have evidence that the gene in question contributes to the development of arteriosclerosis. By exploiting animal models in which levels of implicated factors can be manipulated (genetically or therapeutically) we may gain insight into the molecular pathways to target selectively for prevention or treatment of chronic rejection in humans.

Selected Abbreviations and Acronyms

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Footnotes