Eosinophilia as an early indicator ofpancreatic allograft rejection

Pancreas transplantation has proven to be aneffective strategy for alleviating the morbidity ofpoorly controlled glucose levels in patients withbrittle type 1 diabetes (1, 2). This has beenparticularly important for patients who experiencerecurrent hypoglycemic unawareness. In vivo mon-itoring for pancreas rejection is problematic. Often,serial measures of serum amylase and lipase areused to monitor for rejection in patients withenteric-drained pancreas exocrine secretions (1, 2),whereas urinary amylase monitoring has been usedin patients with pancreatic exocrine secretionbladder drainage (1, 2). Despite utilizing thesemonitoring strategies, rejection can occur withminimal or even only subtle changes in urinaryamylase, or serum amylase, and lipase values.Graft infiltration with eosinophils has been

described in patients with both kidney (3) andmore rarely with pancreas rejection (4, 5). Often,

tissue eosinophilia is a minor part of graft infil-trating cells. When present, eosinophils arestrongly suggestive of rejection (3–5). The percep-tion is that the presence of eosinophils in the graftis a relatively late event and a poor prognostic sign.The resolution of rejection with treatment results indisappearance of the eosinophils (3–5). The role ofthe eosinophil in the rejection is largely unknown,but may be related to alternate pathways of acuterejection (6).

Materials and methods

The patient is a 26-yr-old white women with a longstanding history of type 1 diabetes associated withnephropathy, retinopathy, and autonomic neurop-athy, of which her gastroparesis was one of themost significant problems. She received a livingdonor kidney transplant from her mother on

Weir MR, Bartlett ST, Drachenberg CB. Eosinophilia as an early indicatorof pancreatic allograft rejection.

Abstract: Monitoring pancreas transplant recipients for rejection is aninexact science. Serial monitoring of urinary amylase has been used forpatients with a bladder-drained pancreas. An increase in serum amylase andlipase has been utilized as an in vivo measure of pancreas rejection in pa-tients with enteric pancreatic exocrine drainage. Decreases in urinaryamylase or increases in serum amylase or lipase, respectively, in these twodifferent types of surgical drainage would prompt a pancreas biopsy forhistologic confirmation of rejection. Herein, we describe the case of anenteric-drained pancreatic transplant recipient who presented with periph-eral eosinophilia at least one month before she developed increases in serumamylase and lipase. A pancreas allograft biopsy indicated eosinophilic acutecellular rejection. Peripheral eosinophilia may be a useful early indicator ofpancreas graft rejection preceding changes in serum pancreatic enzymes byapproximately one month.

Matthew R. Weira, StephenT. Bartlettb and Cinthia B.Drachenbergc

aDivision of Nephrology, University of Maryland

School of Medicine, bDepartment of Medicine,

the Division of Transplantation, University of

Maryland School of Medicine and cDepartment of

Surgery, and the Department of Pathology,

University of Maryland School of Medicine,

Baltimore, MD, USA

Key words: amylase – eosinophilia – lipase –

pancreas graft rejection – pancreas transplant –

rejection

Corresponding author: Matthew R. Weir, MD,

Division of Nephrology, University of Maryland

School of Medicine Medical Center, 22 S. Greene

Street, Room N3W143, Baltimore, MD 21201,

USA.

Tel.: 410 328 5720; fax: 410-328-5685;

e-mail: mweir@medicine.umaryland.edu

Conflict of interest: None.

Accepted for publication 17 February 2011

Clin Transplant 2012: 26: 238–241 DOI: 10.1111/j.1399-0012.2011.01440.xª 2011 John Wiley & Sons A/S.

238

October 17, 2008. She had a stable course post-operatively with a serum creatinine that rangedbetween 1.2 and 1.5 mg/dL. She received Thymo-globulin induction for a weak positive crossmatch,and chronic immunosuppression consisted oftacrolimus and low doses of mycophenolic acid(MPA) and a brief course of corticosteroids. Shehad intercurrent problems with autonomic neu-ropathy which required chronic pain control, andshe also required medications for recurrent nausearelated to her gastroparesis.

On May 10, 2009, she received a deceased donorpancreas transplant. She received Thymoglobulininduction coupled with continuation of her tacrol-imus, MPA, and tapering doses of steroids withremoval at approximately three wk post-transplan-tation. Her clinical course was stable for the nextseveral months with tacrolimus levels rangingbetween 6 and 11 ng/mL (12 h trough). HerMPA dose initially was 360 bid, but was discon-tinued when she became leukopenic and wasdiagnosed with cytomegalovirus disease by serumpolymerase chain reaction in October, 2009. Atthis point, she received a course of intravenousganciclovir followed by po valganciclovir. Her doseof MPA was restarted at 360 mg bid after com-pleting ganciclovir therapy. She continued to havelow-grade leukopenia in the 1.5–3.0 range. HerMPA dose was reduced to 360 mg qd and thensubsequently discontinued to allow her white countto improve.

She was first noted to have eosinophilia onDecember 23, 2009, when her white count was 3.3with an eosinophil percentage of 10% (Fig. 1). Hertacrolimus level was 11.7 ng/mL with an amylaseof 52 and a lipase of 19. Approximately two wklater, her eosinophil percentage had increased to21%, with a white blood cell (WBC) of 4.1.Amylase and lipase remained stable at 59 and 24,respectively, and her serum creatinine was 1.3.However, on January 22, 2010, she was first notedto have amylase and lipase elevations of 80 and154, respectively. Her tacrolimus level was 5.3 ng/mL. Her WBC was 7.1 with 23% eosinophils, andwhen repeated five d later, her WBC was 6.8 with37% eosinophils. Her creatinine remained stable inthe 1.2–1.3 mg/dL range throughout this timeperiod.

On February 2, her amylase rose to 435 andlipase to 926, respectively, and she was admitted tothe hospital for a pancreas graft biopsy. Thisbiopsy demonstrated acute cellular rejection withsheets of eosinophils (Fig. 2). She was treated with1 g methylprednisolone x2, and was switched toThymoglobulin. Her CD3 counts did not fallbelow 100 cells/cc. She was subsequently given a

dose of anti CD-52 (alemtuzumab). She wasdischarged on tacrolimus with a goal of 9–10 ng/mL (12 h trough), prednisone 20 mg qd, and MPA360 bid, along with chemoprophylaxis (valganci-clovir, clotrimazole, and trimethoprim-sulfameth-oxazole) and other medications for control of herpain, neuropathy, and gastroparesis. A follow-upbiopsy on February 18, 2010 demonstrated reso-lution of the graft eosinophil infiltration (Fig. 3).

Results

Subsequent to treatment, her eosinophilia disap-peared. On February 15, 2010, her WBC was 5.2with an eosinophil percentage of 3%. By February

Fig. 1. Depicts the time course of % eosinophils and serumlipase. Arrow indicates timing of anti-rejection therapy.

Fig. 2. An adequate pancreas biopsy showed extensive, con-fluent septal and also acinar inflammation predominantlycomposed of eosinophils. Also present were scattered lym-phocytes, macrophages and occasional plasma cells. Acinarinflammation, venulitis, ductitis and intimal arteritis were alsopresent (moderate acute cell-mediated rejection or Banff II).

Eosinophilia and pancreas graft rejection

239

25, her WBC was 7.2 with an eosinophil percent of1%. Amylase and lipase continued to fall and weredown to 86 and 177, respectively, by February 25.She has remained clinically stable subsequently,with lipase levels <20, and eosinophil percent of1%.

Discussion

Prominent eosinophilic infiltrates are an unusualfinding in the pancreas. It has been described inpancreatic allograft rejection, pancreatic pseudo-cyst, lymphoplasmacytic sclerosing pancreatitis,inflammatory myofibroblastic tumor, and histocy-tosis (7). It also may occur in autoimmunepancreatitis (8). Theoretically, it could also occurin drug-induced allergic pancreatitis (9). Moreover,we could not find any evidence in the literature thatperipheral eosinophilia was predictive of eosino-philic pancreatitis and graft rejection.Eosinophilic infiltrates have been described with

graft rejection, both in the kidney (3), and in thepancreas (4, 5). The purpose of the infiltrates isunknown. However, some have suggested thateosinophils could be effectors of rejection whenmore classical pathways of rejection are inhibited,and when T helper 2 (Th2) cells dominate theimmune response (6) (Fig. 4). In experimentalmodels, classic mechanisms of rejection includedelayed-type hypersensitivity by T helper 1 (Th1)CD4+ T cells (10) or by cytotoxic CD8+ Tcells(11). Alternate mechanisms of rejection can be

mediated by the CD95 Fas Ligand dependentcytotoxicity of CD4+ T cells (12) and eosinophilicinflammation elicited by Th2-type CD4+ T cellsvia interleukin (IL)-5 (13, 14). IL-5 is a centralmediator of eosinophil action, and regulates dif-ferentiation, growth, activation and extravasationinto tissues (15). Conceivably, the IL-5/eosinophilpathway of rejection could be more prominent ifother rejection pathways such as that mediated byTh1 CD4+ Tcells, CD8+ T cells, or CD95 (FasLigand) are dampened or even blocked. Themajority of models of eosinophilic graft rejectionlack CD8+ T cell activity (16, 17). Thus, in theabsence of CD8+ T cell activity, the defaultresponse could be a Th2-type alloimmune responsemediated by the IL-5-eosinophil pathway.

It is likely that the immune response in the graftwas substantial and more Th2 directed such thatIL-5 and other chemokines/cytokines recruitedeosinophils to such an extent that they wereevident in the serum weeks before evidence ofgraft infiltration and injury manifested by therelease of amylase and lipase into the systemiccirculation. As such, monitoring the WBC differ-ential may be a useful consideration to evaluate foralternate pathways of rejection in pancreatictransplant recipients.

Acknowledgements

We thank Tia A. Paul, University of Maryland School ofMedicine, Baltimore, MD, for expert secretarial support.

Fig. 3. A repeat biopsy after treatment of acute rejectiondemonstrated disappearance of the eosinophilic infiltrates withmild residual mixed infiltrates (mononuclear and neutrophilic)predominantly within the acini (Banff ‘‘indeterminate’’ grade).Vascular inflammation has resolved and the ducts showedreparative changes. C4d stain was negative in both biopsies.

Fig. 4. This figure depicts classic and alternative rejectionpathways. Classic pathways include cytolytic (CD8+) cells anddelayed-type hypersensitivity, which incorporates T helper 1(Th1) CD4+ cells which activate macrophages (MP) viainterferon-gamma (IF-gamma). Alternate pathways includeFas Ligand-dependent cytotoxicity with CD8+ cytolytic cells,and eosinophilic attack mediated by T helper 2 (Th2) CD4+

cells via interleukin-5 (IL-5) production.

Weir et al.

240

References

1. Bartlett ST, Schweitzer EJ, Johnson LB et al.Equivalent success of simultaneous pancreas kidney andsolitary pancreas transplantation. A prospective trial oftacrolimus immunosuppression with percutaneous biopsy.Ann Surg 1996: 224: 440.

2. Sutherland DE, Gruessner RW, Dunn DL et al.Lessons learned from more than 1,000 pancreas trans-plants at a single institution. Ann Surg 2001: 233: 463.

3. Weir MR, Hall-Craggs M, Shen SY et al. The prog-nostic value of the eosinophil in acute renal allograftrejection. Transplantation 1986: 41: 709.

4. Drachenberg CB, Papadimitriou JC, Klassen DKet al. Evaluation of pancreas transplant needle biopsy:reproducibility and revision of histologic grading system.Transplantation 1997: 63: 1579.

5. Drachenberg CB, Odorico J, Demetris AJ et al. Banffschema for grading pancreas allograft rejection: workingproposal by a multi-disciplinary international consensuspanel. Am J Transplant 2008: 8: 1237.

6. Goldman M, Le MA, Braun M, Flamand V, Abram-

owicz D. A role for eosinophils in transplant rejection.Trends Immunol 2001: 22: 247.

7. Abraham SC, Leach S, Yeo CJ et al. Eosinophilic pan-creatitis and increased eosinophils in the pancreas. Am JSurg Pathol 2003: 27: 334.

8. Sah RP, Pannala R, Zhang L, Graham RP, Sugumar

A, Chari ST. Eosinophilia and allergic disorders inautoimmune pancreatitis. Am J Gastroenterol 2010: 105:2485.

9. Trivedi CD, Pitchumoni CS. Drug-induced pancreatitis:an update. J Clin Gastroenterol 2005: 39: 709.

10. Krieger NR, Yin DP, Fathman CG. CD4+ but notCD8+ cells are essential for allorejection. J Exp Med1996: 184: 2013.

11. Clayberger C. Cytolytic molecules in rejection. CurrOpin Organ Transplant 2009: 14: 30.

12. Chavez-Galan L, renas-Del Angel MC, Zenteno E,Chavez R, Lascurain R. Cell death mechanisms inducedby cytotoxic lymphocytes. Cell Mol Immunol 2009: 6: 15.

13. Chan SY, DeBruyne LA, Goodman RE, Eichwald EJ,Bishop DK. In vivo depletion of CD8+ T cells results inTh2 cytokine production and alternate mechanisms ofallograft rejection. Transplantation 1995: 59: 1155.

14. Matesic D, Valujskikh A, Pearlman E, Higgins AW,Gilliam AC, Heeger PS. Type 2 immune deviation hasdifferential effects on alloreactive CD4+ and CD8+ Tcells. J Immunol 1998: 161: 5236.

15. Kay AB, Barata L, Meng Q, Durham SR, Ying S.Eosinophils and eosinophil-associated cytokines in allergicinflammation. Int Arch Allergy Immunol 1997: 113: 196.

16. Braun MY, Desalle F, Le MA et al. IL-5 and eosin-ophils mediate the rejection of fully histoincompatiblevascularized cardiac allografts: regulatory role of allore-active CD8(+) T lymphocytes and IFN-gamma. Eur JImmunol 2000: 30: 1290.

17. Foucras G, Coudert JD, Coureau C, Guery JC.Dendritic cells prime in vivo alloreactive CD4 T lympho-cytes toward type 2 cytokine- and TGF-beta-producingcells in the absence of CD8 T cell activation. J Immunol2000: 165: 4994.

Eosinophilia and pancreas graft rejection

241