THE IMMUNE RESPONSE FOLLOWING SMALL BOWEL TRANSPLANTATION

I. An unusual pattern of cytokine expression¹

Abstract

Acute cell mediated graft rejection is frequently associated with an immune response dominated by cytokines like IL-2 and IFNγ. While small bowel grafts are rejected acutely, there is little information on the type of immune response generated following transplantation and, in particular, whether the cytokine profile resembles that seen during the rejection of other solid organ grafts. In this paper we compare the expression of cytokines in isolated gut tissue following experimental small bowel transplantation with that in heart grafts. Heterotopic small bowel (n=32) and cardiac (n=32) transplants were performed using the following rat strain combinations: syngeneic Lewis (Lew) > Lew (n=8), blood group D Agouti (DA) > DA (n=8) and allogeneic Lew > DA (n=8), DA > Lew (n=8). Two rats from each group were sacrificed at 1, 3, 5, or 7 days after transplantation. RNA was prepared separately from gut wall, after removing the Peyer's patches (PPs) and mesenteric lymph nodes (MLNs) and from heart. Cytokine (IL-1α, IL-2, IL-4, IL-6, IL-10 and IFNγ) transcripts were analyzed using semiquantitative RT-PCR. Most notably, transcripts of only a single cytokine, IFNγ, became progressively elevated with time in the rejecting small bowel grafts. This is in marked contrast to the findings presented here for rat cardiac grafts in which transcripts kf all cytokines tested show an increase with rejection. This significant and steady increase in IFNγ expression occurred before there was any clinical or histological evidence of rejection. These data demonstrate that the mechanisms of rejection in small bowel and other solid organ grafts are likely to be different. Further, the unique rise in IFNγ expression in the gut wall may be a valuable and early indicator of graft rejection.

Following transplantation of a variety of solid organs, an immune response develops that is dominated by cytokines that are normally associated with Th1 cells ^{(1, 2)}. There are, however, notable exceptions to this finding. For instance, high levels of interleukin (IL)*-4 and/or IL-5 have been observed in liver grafts undergoing acute graft rejection ^{(3, 4)} and exnograft rejection is associated with a Th2 cytokine profile ⁽⁵⁾. Very little is known about rejection of small bowel grafts except that the associated immune response appears to be particularly aggressive and difficult to control with conventional immunosuppressive drugs ^{(6, 7)}.

Small bowel transplants are unique among organ grafts with respect to the load of associated lymphoid tissue and it is clearly important to consider the immune response generated within such tissue. However, it is critical that in any attempt to dissect the mechanisms of small bowel graft rejection, demarcation between events occurring within the gut itself from those in the lymphoid tissue be made. In a first attempt to dissect the immunological mechanisms of small bowel transplant rejection, we have compared cytokine gene expression in rat small bowel allografts with that in rat cardiac allografts. Material was obtained at varying time points after grafting in two strain combinations. Peyer's patches (PPs) and mesenteric lymph nodes (MLNs) were removed from small bowel transplants before analysis was performed. Transcripts for a variety of cytokines (IL-1α, IL-2, IL-4, IL-6, IL-10, and IFNγ) were analyzed using a semiquantitative RT-PCR. We demonstrate that an unusual pattern of cytokine expression is observed in small bowel grafts, the only cytokine showing significant and progressive elevation with time after transplantation being IFNγ.

MATERIALS AND METHODS

Animals. Inbred Lewis (RT1¹, Lew) and blood group D Agouti(RT1^av1, DA) rat strains were used which differ at major and minor histocompatibility loci. They were obtained from Harlan Olac Ltd, Bicester, UK, and maintained in the Biomedical Services Unit at the John Radcliffe Hospital, Oxford.

Transplant model. Heterotopic small bowel transplants were performed using standard microvascular techniques in both syngeneic and fully allogeneic combinations ⁽⁸⁾. The jejunum and ileum, and accompanying MLNs were transplanted as an accessory heterotopic graft, with a venous anastomosis to the inferior vena cava and an arterial anastomosis to the infrarenal aorta. The ends of the transplanted bowel were exteriorized as stomata in the form of a Thiry-Vella fistula and the recipient's intestine was left intact.

The cardiac transplants were performed heterotopically as described by Ono and Lindsay ⁽⁹⁾. The heart graft was anastomosed to the recipient inferior vena cava and infrarenal aorta.

Experimental groups. Small bowel (n=32) and cardiac (n=32) transplants were performed in separate animals using the following rat strain combinations: syngeneic Lew > Lew (n=8), DA > DA (n=8) and allogeneic Lew > DA (n=8) and DA > Lew (n=8). Two rats from each group were sacrificed on days 1, 3, 5, and 7 after transplantation. After this time rejection had progressed such that material could no longer be recovered. In the animals with small bowel transplants a 4 cm segment of mid-ileal gut sall was processed following removal of the PPs. MLNs were excluded from the sample. The heart grafts were removed whole. Untransplanted small bowel(without PPs or MLNs) and cardiac tissue was also obtained from both Lew (n=2) and DA (n=2) animals.

Diagnosis of rejection. Acute rejection was confirmed in the small bowel allograft model by the presence of an abdominal mass^{(8, 10)} and by hematoxylin and eosin staining of gut wall sections to identify established histological changes seen during rejection ⁽¹¹⁾. The heart allografts were monitored daily and were found to be beating normally on day 5 and weakly by day 7.

Preparation of RNA and semiquantitative polymerase chain reaction. The gut wall and heart tissue were separately homogenized in guanidine thiocyanate solution using a Janke and Kunkel Ultraturrax T25 tissue homogenizer (Oxford Laboratories, High Wycombe, UK). Total RNA was prepared as previously described ⁽¹²⁾ and pooled from the 2 animals sacrificed at the same time point in each group. A selection of unpooled RNA samples was analyzed to ensure consistency amongst material. 1 μg of pooled or unpooled RNA was used in the cDNA synthesis with murine molony leukemia virus reverse transcriptase (Gibco-BRL, UK). 1/20th of each resulting cDNA sample was then added to a reaction mixture containing 50 mM KCl, 10 mM Tris HCl (pH 8), between 1.0 and 4.5 mM MgCl₂ (depending on the primer set), 0.01% gelatin, 0.2 mM (each) dNTPs, 1 mM each primer and 2.5 U Taq polymerase (N801-0046; Perkin-Elmer Corp, Hayward, CA) in a total volume of 100 μl for PCR amplification.

Samples were prealiquoted into separate 15 μl volumes and incubated in the DNA thermal cycler (Gene Amp PCR System 9600 Perkin Elmer Corp, Hayward, CA) for a maximum of 35 cycles. Each cycle consisted of 30 sec at 94 °C, 1 min at 55 °C or 60 °C (depending on the primer set), and 30 sec at 72°C. Aliquots (15 μl) of the PCR reaction mix were removed at 5 cycle intervals (20-35 cycles), transferred onto a nylon membrane, probed using aγ-[³²P] ATP endlabeled internal oligonucleotide and analyzed byβ scintillation counting (Wallac OY, Finland) and/or autoradiography. All samples were controlled by RT-PCR using β actin specific primers (15-30 cycles, not shown). A second sample obtained after the maximum number of cycles was analyzed by gel electrophoresis to check integrity and size of PCR product (not shown).

Sense (A) and antisense primers (B) and internal oligonucleotides (I) were synthesized on a 380B DNA synthesizer (Applied Biosystems, Warrington, UK) based on published cDNA sequences for each of the following: IL-1α⁽¹³⁾, IL-2 ⁽¹⁴⁾, IL-4⁽¹⁵⁾, IL-6 ⁽¹⁶⁾, IL-10⁽¹⁷⁾, IFNγ ⁽¹⁸⁾, andβ-actin ⁽¹⁹⁾ control primers. The MgCl₂ concentration and annealing temperature used for each primer pair were 1.25 mM and 55 °C for IL-1α, 1.5 mM and 60 °C for IL-2, 1.25 mM and 55°C for IL-6, 1.5 mM and 55 °C for IL-10, 1.75 mM and 60 °C for IFNγ, and 1.5 mM and 55 °C for β actin. Oligonucleotide sequences were as follows: Equation

Statistics. The changes in expression of IFNγ in the gut wall of the allogeneic and syngeneic models were measured by simple linear regression of counts per minute on days posttransplantation.

RESULTS

Clinical and histological signs of rejection. In syngeneic small bowel transplants no clinical or histological changes were observed at any time. A mass was consistently palpable on day 4 in both allogeneic strain combinations. The first histological changes seen in the allogeneic gut wall appeared between day 5 and 7 with cryptitis and early shortening and blunting of the villi. By day 7 there was a dense lymphocytic and plasma cell infiltration in the lamina propria, muscularis propria, and submucosa, associated with epithelial sloughing.

All the syngeneic cardiac grafts beat consistently well at all the time points after transplantation. However, although allografts were all beating well on the fifth postoperative day beat rates and strength rapidly declined thereafter and by day 7 were often barely detectable. Rejection in this model is usually clinically complete between day 7 and 9 and histologically it is established by day 5 ⁽²⁰⁾.

Cytokine expression in heart tissue(Figs. 1-5, Table 1). In heart allografts, all measurable cytokine transcripts increased after grafting and the peak level of IL-2 (Fig. 1A, Table 1), IL-10 (Fig. 2A, Table 1) and IFNγ(Fig. 3A and 3B, Table 1) transcripts was reached on day 5. IL-2 transcripts were not found or only barely detectable in either the syngeneic grafts or untransplanted heart tissue, whereas all other cytokines could be detected at low level in both situations. Both IL-6(Fig. 4A, Table 1) and IL-10 exhibited proinflammatory cytokine characteristics in that there was a early increase of expression not only in allogeneic grafts, but also in syngeneic grafts. A second wave of IL-6 expression occurred after the time that heartbeat had begun to decline. IL-1α was more strongly expressed in allogeneic grafts than in isografts (Fig. 5A, Table 1). IL-4 transcripts could not be detected in any tissue at any time.

Cytokine expression in small bowel(Figs. 1-5, Table 1). IL-1α (Fig. 5B, Table 1), IL-2(Fig. 1B, Table 1), IL-6(Fig. 4B, Table 1), IL-10(Fig. 2B, Table 1) and IFNγ(Fig. 3C and 3D, Table 1) transcripts were all expressed in normal gut wall. IL-1α, IL-6 and IL-10 transcripts were slightly elevated in syngeneic grafts. Only in the case of IL-10 and IFNγ was there a more substantial increase of transcripts in the allogeneic than in the syngeneic grafts. Of most note was the steady rise in IFNγ transcripts with time following grafting. To evaluate the significance of this finding, data obtained for IFNγ were further analyzed by β scintillation counting of the ³²P-labeled dot blots(Fig. 3D). Simple linear regression analysis of the data indicated that in the small bowel grafts, the rise in transcript level was significant by day 3 in both DA>Lew (P=0.02) and Lew>DA(P=0.035) strain combinations. In contrast, no significant change in IFNγ transcripts was observed with time in their syngeneic counterparts(DA>DA, P=0.94; Lew>Lew, P=0.17). IL-4 transcripts were not detected in any tissue at any time.

DISCUSSION

The aim of this study was to provide an insight of the immune response to small bowel allografts by analyzing cytokine expression. Using a simple semi-quantitative RT-PCR we have been able to demonstrate that small bowel allografts differ dramatically from cardiac allografts not only in the profile, but also in the kinetics, of cytokine expression. A clear Th0/Th1 pattern of cytokine transcripts, similar to that which was seen in the cardiac transplants shown here and which has been noted in several other transplants models, was not observed. Further, there was no evidence of a dominant Th2 response suggesting perhaps a unique mechanism(s) of small bowel graft rejection.

It is perhaps not surprising that all the cytokines tested in this study, including IL-2, were readily detectable in both normal gut wall and syngeneic transplants since the gut wall contains leukocytes and is continuously exposed to foreign antigen. It is, however, surprising that in the gut wall, only the pattern of IFNγ expression correlated with the rejection process. The low and unrelated expression of IL-2 in the gut wall suggest that the cells responsible for the production of IFNγ do not secrete IL-2 and therefore may not be conventional activated T cells.

In the only other reported study of a similar nature, McDiarmid et al.⁽²¹⁾ demonstrated an elevation of several cytokine transcripts, including those of both IL-2 and IFNγ. However in this study, no attempt was made to remove PPs from the analyzed material and so it is not possible to define the source of each transcript. Although it clearly is important to analyze the expression of cytokines within local lymphoid tissue, more information may be gleaned about the rejection process if gut and lymphoid compartments are analyzed separately. Indeed, studies nearing completion in this laboratory have clearly demonstrated a very different pattern and kinetic of cytokine expression in PPs and MLNs (Toogood et al., unpublished data). Analysis of unseparated compartments masks these important differences.

Since the isolated increase in IFNγ expression during rejection is apparently unique to the gut wall compartment of the small bowel graft, it seems likely that there may be an unusual cell type in the gut wall involved in its production which may be of donor or recipient origin. The marked levels of IFNγ expression in the normal gut wall suggest that, at least in part, the IFNγ producing cells may be of recipient origin. Intraepithelial lymphocytes are known to have NK cell activity⁽²²⁾ and produce IFNγ^{(23, 24)} and therefore, may be stimulated to produce IFNγ during the rejection. Cell culture studies show that NK cells secrete IFNγ when stimulated by a number of different cytokines, including IL-2, IL-7, and IL-12 ⁽²⁵⁾. IL-12, which is produced by macrophages and was previously known as NK stimulatory factor, appears to be the most effective of these and is perhaps the most likely candidate responsible for the elevation of IFNγ shown here, since the levels of IL-2 appear to be low in the gut wall and there is a known macrophage infiltrate in the muscularis mucosa and submucosa 3 days after transplantation ⁽²⁶⁾.

There is an infiltration of host lymphocytes into the lamina propria and intraepithelial compartments during the first few days after transplantation⁽²⁷⁾. This infiltrate may predominate in cytotoxic T cells and NK cells, which can produce IFNγ, although T helper cells also produce IFNγ and are likely to be present. The lack of elevation in IL-2 transcript, however, would suggest that T helper cells may not be responsible for the observed IFNγ production. Nevertheless, whichever cells are responsible for producing IFNγ, they are likely to be increasingly stimulated by products from the infiltrating macrophages and/or T cells.

The progressive increase in IFNγ expression correlates well with the known early increase in MHC class II expression seen on enterocytes⁽²⁸⁾ and the host infiltration of lymphocytes into the lamina propria and intraepithelial compartment during the first few days after transplantation. Further, IFNγ reduces the barrier function of the gut wall ⁽²⁹⁾ and bacterial translocation is known to increase during rejection.

The early peak in the level of IL-6 transcripts and the overall raised levels of expression in transplanted tissue as compared with untransplanted tissue presumably reflects the inflammatory process in the graft resulting from the trauma of surgery. In the allogeneic strains there was a second peak of IL-6 expression on day 5 to 7 in the gut wall, which reflects the inflammatory process taking place as a result of the rejection, a finding very similar to that seen in the heart model. However, in both the heart and gut wall allografts the increase in IL-6 expression occurred only after there was clinical and histological evidence of tissue destruction. The increased expression of IL-1α in both the heart and gut wall allografts correlates with the acute turnover of inflammatory cells that takes place during the rejection process. The generally elevated levels of IL-10 transcript in allogeneic grafts may also relate to inflammation, but bore no real relationship to the progressing immune response.

Acute rejection in clinical small bowel transplantation continues to be a significant cause of morbidity. Accurate and early diagnosis of rejection is often particularly difficult to achieve in small bowel transplantation. There are no simple biochemical markers to follow, the clinical signs are vague and nonspecific and the histological changes are late and patchy⁽³⁰⁾. Therefore, a reliable and accurate marker of rejection is particularly needed in small bowel transplantation. As only very small amounts of tissue are required for this technique and stomal biopsies do not result in any discomfort to the patient, the level of IFNγ expression in the gut wall could easily be measured following clinical transplantation. Basal levels of IFNγ expression could be measured in pretransplant biopsies and increases in the level could provide a novel early marker of rejection from a site that is easily accessible in clinical small bowel transplantation.

In conclusion, it is apparent that there are significant immunological differences between the gut wall compartment of a small bowel transplant and other vascularized allografts. Further analysis of the source of the IFNγ may lead to a better understanding of the rejection process that takes place in the small bowel transplant.

Acknowledgments. We thank Heather Cordell for her advice on statistics and Mr. N. E. Dudley for his help with this work.

Figure 1: Semiquantitative dot blot analysis of IL-2 expression in heart and gut wall. IL-2 primers were used to amplify cDNA samples from heart (A) and gut wall (B) in syngeneic (days 1, 3, 5, 7 posttransplantation), allogeneic (days 1, 3, 5, 7 posttransplantation), and untransplanted tissue(normal Lew and DA). Aliquots of PCR mix were removed at 5 cycle intervals^(20-35), transferred onto a nylon membrane, probed using γ-[³²P] ATP end-labeled internal oligonucleotide and analyzed by autoradiography. (-), no cDNA (negative control); (+), cDNA from activated lymphocytes (positive control).

Figure 2: Semiquantitative dot blot analysis of IL-10 expression in heart and gut wall. Analysis was performed using IL-10 primers with material from heart (A) or gut wall (B) as described for Figure 1.

Figure 3: Semiquantitative analysis of IFNγ expression in heart and gut wall. (A and C) Dot blot analysis was performed using IFNγ primers on heart (A) or gut wall (C) as described for Figure 1.(B and D) Results of PCR for heart (B) or gut wall (D) were quantified byβ scintillation counting. The counts per minute relate to the 30 cycle samples (similar results were obtained at other cycles). Day 0, normal untransplanted tissue of DA (▵ and ▪) or Lew (▴ and □).

Figure 4: Semiquantitative dot blot analysis of IL-6 expression in heart and gut wall. Analysis was performed using IL-6 primers with material from heart (A) or gut wall (B) as described for Figure 1.

Figure 5: Semiquantitative dot blot analysis of IL-1α expression in heart and gut wall. Analysis was performed using IL-1α primers with material from heart (A) or gut wall (B) as described forFigure 1.