Operational Tolerance in Intestinal Transplantation

Abstract

By presenting the first case report of true operational tolerance in an intestinal transplant patient, we aim to demonstrate that tolerance is possible in a field which has been hampered by sub-optimal outcomes. While operational tolerance has been achieved in liver and kidney transplantation, and some intestinal transplant patients have been able to decrease immunosuppression, this is the first instance of true operational tolerance after complete cessation of immunosuppression.

A patient received a deceased-donor small intestinal and colon allograft with standard immunosuppressive treatment, achieving excellent graft function after overcoming a graft-versus-host-disease episode 5 months post-transplant. Four years later, against medical advice, the patient discontinued all immunosuppression. During follow-up visits 2 and 3 years after cessation of immunosuppression, the patient exhibited normal graft function with full enteral autonomy and without histological or endoscopic signs of rejection. Mechanistic analysis demonstrated immune competence against third party antigen, with in vitro evidence of donor-specific hyporesponsiveness in the absence of donor macrochimerism.

This proof of principle case can stimulate future mechanistic studies on diagnostic and therapeutic strategies, for example cellular therapy trials, that can lead to minimization or elimination of immunosuppression and hopefully help revitalize the field of intestinal transplantation.

Case Report

Introduction

Intestinal transplants (ITx) are declining due to poorer rates of success than other solid organs given the gut’s complex immunologic milieu, associated with high rates of rejection, graft-versus-host-disease (GVHD), and frequent over-immunosuppression.¹ While nearly 200 ITx were performed in the United States in 2007, current trends suggest less than 100 ITx will be performed in 2019.² The achievement of ITx operational tolerance could transform the current clinical landscape, allowing broader application of this lifesaving treatment.

Initial Postoperative Course

A 14-year-old male with chronic intestinal failure secondary to megacystis microcolon intestinal hypoperistalsis (Berdon’s Syndrome) was accepted for ITx due to progressive inanition, disability due to severe abdominal distention, and total parenteral nutrition (TPN) dependence.

His transplant included small bowel enterectomy and subtotal colectomy, subtotal gastrectomy for megalogastria, and implantation of an allograft including jejunoileum and right colon from a blood group identical 17-year-old brain-dead male. Both donor and recipient were sero-negative for cytomegalovirus (CMV), and lymphocytotoxic T and B cell crossmatches were negative. Allograft implantation was orthotopic, between donor and recipient superior mesenteric arteries and veins. Cold ischemia time was 5 hours and 19 minutes; warm ischemia time 19 minutes. Enteral continuity between donor and recipient was achieved proximally via both a gastro- and duodenojejunostomy and distally via colo-colostomy. A jejunal feeding tube and loop ileostomy were placed as standard after ITx.

Donor and recipient HLA types were completely mismatched (all 6 ABDR antigens). Induction immunosuppression consisted of basiliximab (20mg days 0 and 4) and a standard methylprednisolone taper. Maintenance immunosuppression initially included tacrolimus (mean trough levels 25.2ng/mL for first 4 weeks, then tapering), rapamycin (mean trough levels 7.8ng/mL for first 4 weeks), and prednisone (20mg per day). Rapamycin was discontinued after 7 weeks due to intolerance and replaced by mycophenolate mofetil (500mg twice per day). The patient received standard infectious prophylaxis including intravenous CMV hyperimmune globulin and trimethoprim-sulfamethoxazole. Early endoscopic and histologic assessment of the graft demonstrated mild ischemia-reperfusion injury-related changes resolving in the first week. The patient tolerated rapid progression to full enteral nutrition by the time of discharge 3 weeks post-transplant.

Graft-versus-Host-Disease Episode

Post-transplant follow-up showed a normal, healthy allograft over the course of the first 5 months post-transplant, at which time the patient successfully underwent loop ileostomy closure. During this admission, he complained of rectal discomfort and a skin rash on his fingers. A punch skin biopsy demonstrated evidence of subacute lichenoid GVHD and a pan-endoscopy showed erythematous native colon, with histology results pointing to grade 3 GVHD. Chimerism testing was performed on genomic DNA isolated from whole blood using amplification of 8 STR/VNTR loci containing tandemly repeated sequences as specified in Supplementary Materials and Methods. At 7 out of 8 loci, only markers that matched the recipient were detected. At 1 of the 8 loci, possible chimerism of 2% was detected. However, since this was not confirmed at the other loci, which would have been expected in the case of chimerism, the conclusion was reached that no chimerism was present; although it cannot be categorically ruled out given the finding in 1 of the 8 loci.

Tacrolimus dose was reduced by 50% in an initial effort to achieve resolution of GVHD via native immune reconstitution. The patient was re-admitted 2 months after the ileostomy takedown (7 months post-transplant) with increased stool output, flatulence, urgency, and unintentional 5kg weight loss. Endoscopic and histologic evaluation of the native colon demonstrated worsening grade 3 GVHD. He was treated with a 7-day course of rabbit anti-thymocyte globulin (1.5mg/kg bodyweight per day) and high dose steroids (prednisolone 2mg/kg bodyweight per day) followed by taper with complete clinical, endoscopic, and histologic resolution of GVHD 7 months later (14 months post-transplant). Subsequent maintenance immunosuppression included 3mg of prednisone daily, tacrolimus to maintain a trough level of 5–8ng/mL, and mycophenolate mofetil to maintain a trough mycophenolic acid level of 1–2mcg/mL.

Weaning Off Immunosuppression by Choice

Following GVHD resolution, his clinical course was unremarkable with excellent growth and weight gain, normal labs, and normal graft endoscopic and biopsy findings, excepting intermittent mild reflux esophagitis. His aforementioned maintenance immunosuppressive regimen continued. He also received lansoprazole (30mg daily), levothyroxine (27.5mcg daily), and somatotropin (3.2mg six times per week) for treatment of comorbid conditions (mild esophagitis, hypothyroidism, and growth hormone deficiency).

During a routine annual follow up visit 3 years and 7 months post-transplant, the patient was found to have a positive de novo donor specific antibody (DSA) against the donor antigen DQ4 (titer >1:1024, peak mean fluorescent intensity (pMFI)=12682, C1q strongly positive), in the setting of likely non-adherence with his maintenance immunosuppressive regimen. He was then scheduled to receive treatment with high dose intravenous immunoglobulin (2g/kg bodyweight per treatment), as repeat DSA studies showed persistently unchanged high DSA levels; however, he refused treatment and completely stopped taking any maintenance immunosuppressive medications (starting circa 4 years post-transplant). Despite the positive DSA, C4d immunohistochemical (IHC) staining of his allograft was negative. Moreover, he continued to be in good health with excellent graft function and no signs of complications or rejection. Thanks to the ITx, he was able to graduate from high school and transition to college.

During follow-up visits 2 and 3 years later (circa 6 years and 7 years post-transplant, respectively) he confirmed complete non-adherence with all immunosuppressive medications, which was corroborated by undetectable tacrolimus and mycophenolic acid blood levels. Moreover, with the exception of two brief, spontaneously resolving episodes of watery diarrhea and nausea, he reported a normal stool pattern, and no abdominal pain, bloating, feeding intolerance, or problems maintaining weight. Indeed, lab tests as well as pan-endoscopic assessments of both his allograft and native remnant colon confirmed good allograft health without endoscopic or histologic evidence of rejection or recurrent GVHD (Figure 1), suggesting a state of operational tolerance. At both time points, chimerism testing was again performed. Specifically, at the first time point a blood sample and at the second time point a blood sample and CD3+, CD33+, CD19+, and CD56+ sorted cells were tested at 8 loci. At both time points, only markers that matched the patient were detected at all 8 loci, suggesting absence of donor chimerism. The patient has continued to stay off all immunosuppression and remained in good health more than 9 years post-ITx.

Figure 1: Endoscopic and histologic assessment of intestinal allograft and native colon confirming good allograft health and absence of GVHD in operationally tolerant intestinal transplant recipient.

Upper row: Representative endoscopic images of allograft ileum and colon as well as native colon seven years and two months after transplant (from the most recent colonoscopy, 3 years after stopping immunosuppression). Lower row: Representative images of histologic H&E stains at original magnifications of 20x with inserts at 40x of allograft ileum and colon as well as native colon seven years and two months after transplant. The photomicrographs were captured from scanned slides using nanozoomer 2.0RS digital slight scanner (Hamamatsu, Iwata City, Japan) and analyzed using NDP.view2 software (Hamamatsu, Iwata City, Japan).

Immunological Characterization of Operational Tolerance

Analysis of the patient’s peripheral blood obtained during the two last visits showed his white blood cell, neutrophil, and lymphocyte subsets including CD19+ B cells, CD3+ T cells, CD4+ T cells, and CD8+ T cells to be normal³. Further analysis of CD4+ and CD8+ T cells revealed normal ranges for frequencies of naïve, effector, and memory T cell subsets (Supplementary Figure S1A).³ Further work-up showed the presence of robust populations of regulatory, FoxP3-expressing T cells (Tregs) in his peripheral blood and intestinal allograft lamina propria (Supplementary Figures S1B, S2A–C). With respect to the graft, we found that the mean % of FoxP3+ of CD4+ cells at the two operational tolerant patient time points was 32.6%, versus 20.9% in 9 demographically matched controls. Of note, repeat DSA assays showed persistently unchanged high anti-DQ4 DSA levels, akin to what has been reported in cases of operationally tolerant liver transplant recipients.⁴ However, C4d IHC stains on ileal allograft and native colon were consistently negative.

Functional T cell assays revealed that he was otherwise immunocompetent with normal anti-viral CD3+ T cell reactivity against CMV antigen pp65 and Epstein-Barr-Virus (EBV) lysate (Figures 2A, 2B). Functional assessment of alloreactivity and antigen presentation with mixed lymphocyte culture and antigen presentation assays,^5–7 demonstrated that he had evidence of donor-specific hyporesponsiveness at both a T cell and antigen presentation level. Specifically, he showed a hyporesponsive CD154-expression response of his CD8+CD45RO+ memory T cells against donor-HLA matched stimulation when compared to stimulation with HLA mismatched third-party (Figure 3A). Moreover, there was reduced donor antigen presentation by the recipient’s B cells and monocytes when compared to presentation of HLA mismatched third-party alloantigen (Figures 3B, 3C).

Figure 2: Immunocompetence and cell-mediated antiviral immunity to CMV and EBV in operationally tolerant intestinal transplant recipient.

Recipient peripheral blood T cells were cultured alone, or with CMV-pp65 antigen (A), EBV viral lysate (B), or the mitogens, PHA or PMA (A and B), and CD154+ T cells measured in each culture condition. For each antigenic stimulant, CD154+ T cell frequencies approximate corresponding mean frequencies in 29–40 healthy adults, or 37–145 recipients of liver or intestine transplants (please see Table in the supplementary materials and methods section).

Figure 3: Donor-specific hyporesponsiveness in operationally tolerant intestinal transplant recipient.

(A) Flow cytometry scatterplots show CD154+ T cytotoxic memory cells (TcM) induced by stimulating recipient peripheral blood leukocytes (PBL) for 16 hours with donor or HLA-mismatched reference PBL as described previously.⁵ Also shown is background expression without stimulation. The ratio of donor- and reference-induced CD154+ TcM or the index of rejection was 0.5 in this recipient consistent with decreased donor-specific CD154+ TcM relative to those induced by reference. A ratio < 1.1 predicts low risk of rejection with specificity of 80%.²⁰ The p value for the Poisson test between background and donor-induced CD154+T-cytotoxic memory cells was 0.0002 and the p value for the Poisson test between background and third party reference was 2.20E-16.

(B and C) Flow cytometry scatterplots show frequencies of B cells (B) and monocytes (C) that present fluorochrome-labeled antigenic lysate prepared from donor and reference PBL, as described previously referenced.^6,7 The antigen presenting index (API) is the ratio of B cell or monocyte frequencies that present donor- and reference antigen. The B cell API is 14.2%/16.3% or 0.87 and below the rejection threshold of 1.14. The monocyte API is 0.16 and below the rejection threshold of 1.2 (see supplementary materials and methods section).

Discussion

We present the first reported case of operational tolerance after ITx, defined by Ashton-Chess et al. as long-term and stable normal organ function in an immunocompetent host without evidence of rejection in the absence of all immunosuppression for at least one year after deceased-donor transplant.⁸

To put our case into context, it is of note that operational tolerance in human solid organ transplantation remains extremely rare, and has been almost exclusively described in recipients of liver^9–11 and kidney¹² transplants. For example, a report found that of 461 attempts to wean liver transplant recipients off immunosuppression,¹⁰ only 100 (22%) were successful 1-year out, and that in kidney transplantation only 100 cases of tolerance had ever been reported (20,000 kidney transplants are performed per year).¹² At the time of said report there had been one reported case each in lung and heart transplantation, and none in pancreas or intestine.¹³

Moreover, operational tolerance cases are distinguishable into three categories¹³ (1) spontaneous tolerance due to immunosuppression non-compliance; (2) planned weaning under clinical supervision; and (3) therapeutically driven tolerance via targeted tolerogenic protocols.

For categories (1) and (2), most known cases are liver transplant cases, not surprising given the 22% weaning success rate noted above. For pediatric liver recipients, it may even be possible to achieve tolerance in 60% of cases.⁹ While there has been one previous claim of “prope” tolerance in a recipient of a living donor ITx, it was not complete operational tolerance as this patient was not totally off immunosuppression, had documented rejection, and underwent no immunologic analysis.¹⁴ In contrast, our case demonstrated evidence of clinical organ acceptance (no rejection), immunocompetence to third party antigen, and donor-specific hyporesponsiveness even five years after stopping all immunosuppression.

For category (3) – tolerance induction trials – there is some precedence in ITx. Nearly two decades ago, attempts were made by Thomas Starzl’s group to wean solid organ transplant recipients, including 11 ITx recipients, off immunosuppression by dose-spacing tacrolimus monotherapy to every other day or longer intervals starting four months post-transplantation.¹⁵ This approach was combined with pre-treatment lymphocyte depletion with rabbit anti-thymocyte globulin before transplantation aimed at exhaustion or deletion of donor-specific T cell clones.¹⁵ While 6 of the ITx cases saw some success, they are not examples of true operational tolerance as patients remained on space-dosed immunosuppression.

The hypotheses underlying the Starzl trials, as well as prior graft acceptance studies, are informative of the mechanisms underlying tolerance. One such factor pertains to graft-versus-host and host-versus-graft responses.^15,16 However, while our patient’s GHVD episode was an indicator of an early graft-versus-host reaction and implied that chimerism may have been present at some point – especially given the sensitivity limitations of the VNTR/STR method of chimerism detection – the lack of macrochimerism in the latter two time points leads us to believe that macrochimerism-associated mechanisms of tolerance induction were not the main drivers.

A more prominent factor for operational tolerance pertains to Treg involvement.^11,17 Although we did not directly confirm the involvement of donor-specific Tregs, our analysis demonstrates a robust population of Tregs in our patient’s graft (Supplementary Figure S2A, S2C), implicating peripheral tolerance via Tregs as a potential mechanism. For instance, the reduced CD154 expression response of memory CD8+ T cells stimulated with donor-HLA matched cells in our patient (Figure 3A) could have been mediated by donor-specific Tregs, which have been shown to inhibit the expression of CD154 on T cells.¹⁸ Moreover, the reduced donor antigen presentation by our patient’s antigen presenting cells (APCs) could also be explained by Treg-mediated mechanisms, as Tregs have been shown to directly interact with APCs and disrupt antigen-presentation to T cells.¹⁹ We acknowledge that the tests used (Figure 3) were developed for monitoring rejection and not as markers for operational tolerance. However, given that the tests measure donor-specific hyporesponsiveness, which is a detectable phenomenon in operational tolerance, the tests can be considered a valid proxy.

Paradoxically, our operationally tolerant recipient developed persistent anti-DQ4 DSA after weaning off immunosuppression, akin to what was seen in the majority of pediatric liver transplant recipients successfully withdrawn from immunosuppression and in adult liver transplant recipients who received Treg-based cell product for tolerance induction.^4,11 Despite strong C1q positivity, there was no evidence of C4d deposition or microvascular inflammation in our patient’s allograft as seen in operationally tolerant pediatric liver transplant recipients.⁴ Based on this, it can be speculated that our tolerant intestinal allograft may show low levels of constitutively expressed HLA Class II antigens comparable to tolerant liver allografts.⁴ The expression pattern dynamic of Class II antigens in tolerant versus rejecting intestinal allografts requires further investigation.

In summary, we present the first known case of true operational tolerance after ITx with stable, long-term good graft function off immunosuppression in an immunocompetent host and with donor-specific hyporesponsiveness. This proof of principle case has the potential to stimulate future mechanistic ITx studies aimed at developing diagnostic and therapeutic strategies, for example Treg trials, that can result in minimization or elimination of immunosuppression and bring some much-needed hope to the field of ITx.

Abbreviations:

APC

Antigen presenting cells

API

antigen presenting index

CMV

Cytomegalovirus

DSA

donor specific antibody (DSA)

EBV

Epstein-Barr-Virus

GVHD

Graft-versus-host-disease

IHC

Immunohistochemistry

ITx

Intestinal transplantation

PBL

peripheral blood leukocytes

pMFI

Peak mean fluorescent intensity

STR/VNTR

Variable number of tandem repeat (VNTR) and short tandem repeat (STR) sequences

TcM

T cytotoxic memory cells

TPN

Total parenteral nutrition

Tregs

Regulatory, FoxP3-expressing T cells