We present a case of disseminated histoplasmosis, complicated by retroperitoneal bleeding and leading to death, in a patient who was receiving systemic immunosuppressive therapy for rheumatoid arthritis and who was enrolled in a gene-therapy trial. This trial was designed to evaluate intraarticular delivery of a tumor necrosis factor
(TNF-
) antagonist, through an adeno-associated virus (AAV) type 2 delivery system, for inflammatory arthritis. The patient's receipt of concurrent anti–TNF-
therapy and other immunosuppressive therapy while she was living in an area where histoplasmosis was endemic was thought to be the most likely explanation for the infection; the evidence presented suggests that this fatal infection was unlikely to have been related to exposure to the agent administered in the gene-therapy trial. This case reinforces the importance of considering infectious complications, such as those from endemic mycoses, in patients receiving treatment with a TNF-
antagonist and the importance of having a well-designed monitoring plan when subjects in a research study become ill. (ClinicalTrials.gov number, NCT00126724
[ClinicalTrials.gov]
.)
inhibitors have represented a dramatic improvement in the treatment of rheumatoid arthritis; they include the two antibodies infliximab and adalimumab, as well as etanercept, a dimeric fusion protein that combines an immunoglobulin domain with a TNF-receptor domain (TNFR:Fc).2,3 Trials of TNF-
inhibitors delivered with the use of gene-therapy methods in patients with rheumatoid arthritis are increasing in number.4 The most common vectors used in models of rheumatoid arthritis are viral vectors such as lentivirus, adenovirus, and AAV.5 AAV vectors have been used in gene-therapy trials for multiple diseases, including cystic fibrosis and Duchenne's muscular dystrophy.4,6 The goal of the trial was to evaluate the safety of a TNF-
inhibitor administered either as intraarticular therapy in patients with inflammatory arthritis who had persistent destructive synovitis in selected joints, despite otherwise effective systemic therapy, or as targeted treatment in patients with monoarticular or oligoarticular arthritis, potentially obviating the need for systemic therapy. The gene-therapy agent that was used is an AAV vector in which the rep and cap genes are replaced with the complementary DNA (cDNA) sequence equivalent to that for etanercept; the vector is then packaged in an AAV2 capsid. The expressed protein inhibits TNF-
, a key mediator of inflammation. Methods
The initial gene-therapy study was approved by the Western Institutional Review Board in Olympia, Washington, and the institutional review boards at Northwestern University in Chicago and Presbyterian Hospital of Dallas, Texas Health Resources. After the patient died, in addition to a routine autopsy, further tests were performed in consultation with the Food and Drug Administration (FDA), the Recombinant DNA Advisory Committee of the National Institutes of Health, the study sponsor (Targeted Genetics), and outside experts. The institution at which the autopsy was performed was not affiliated with the gene-therapy study. The family provided written consent to disclose all data. Special testing was funded by the sponsor as part of the FDA investigation. The sponsor had no role in the preparation of the manuscript or the decision to submit it for publication, and the authors vouch for the completeness and accuracy of the data.
Case Report
A 36-year-old woman with a 15-year history of rheumatoid arthritis presented on July 5, 2007, with fever, chills, fatigue, nausea, vomiting, and abdominal pain. Her initial therapy for rheumatoid arthritis had begun in 1992 with hydroxychloroquine, and from 1994 through 2001 she received various combinations of the following medications: methotrexate, azathioprine, prednisone, and etodolac. In 2002, she was treated with etanercept and then with etanercept and methotrexate. Since 2004, she had been treated with adalimumab (40 mg every other week), methotrexate (20 mg once a week), low-dose prednisone (3 mg once a day), and cortisone injections in several joints, primarily the right knee. Adalimumab therapy was interrupted for several months in late 2006 when the patient underwent foot surgery. Her reason for enrolling in a gene-therapy study to treat rheumatoid arthritis was the presence of persistent knee synovitis. At study screening, the results of tests for tuberculosis, human immunodeficiency virus (HIV) infection, hepatitis B, and hepatitis C were negative. On February 26, 2007, she received the first injection, in the right knee, of the active gene-therapy agent, tgAAC94, which was made up of 5x1013 DNase-resistant particles of the recombinant AAV2 vector plus a cDNA insert designed to express the TNF-
antagonist, consisting of an immunoglobulin domain fused to a TNF-receptor domain. At this time, she continued to receive adalimumab, methotrexate, and prednisone at the doses listed above. On July 2, 2007, the day of the second injection of the same active dose, she reported fatigue of several days' duration and low-grade fever for 1 day. No change in her systemic immunosuppressive medications had been made in the 4-month period before the second injection. She had a temperature of 37.6°C (99.6°F), but her vital signs were otherwise normal. That evening, she had nausea, vomiting, fever, and chills, followed by diarrhea and abdominal pain, particularly in the epigastric region. Because of her continued fever, she was given levofloxacin orally, 500 mg a day, starting on July 5. Five days after the injection, she presented to an emergency room with a temperature of 40°C (104°F), nausea, vomiting, and headache. A chest radiograph, blood chemical tests, and a complete blood count were normal. Blood and urine cultures drawn at this time were ultimately found to be negative. She was discharged with a diagnosis of a viral syndrome; promethazine was prescribed for the nausea. She continued to take the levofloxacin daily.
On July 9, during visits to the physician associated with the gene-therapy trial and to her primary care physician, she reported intermittent fevers, headaches, and vomiting. She had tachycardia (heart rate, >100 beats per minute) but a nontender abdomen. Laboratory tests showed an elevated white-cell count (29,200 per cubic millimeter) and elevated liver-function test results (alanine aminotransferase level, 125 U per liter; aspartate aminotransferase level, 162 U per liter; and alkaline phosphatase level, 139 U per liter) (Table 1). Three days later, on July 12, she was admitted to the hospital with abdominal pain and distention, nausea, vomiting, diarrhea, fever, chills, jaundice, abnormal liver-function test results (alanine aminotransferase level, 147 U per liter; aspartate aminotransferase level, 291 U per liter; and alkaline phosphatase level, 455 U per liter), an elevated white-cell count (17,600 per cubic millimeter), and thrombocytopenia (a decrease in platelets from 149,000 per cubic millimeter to 86,000 per cubic millimeter). The combined results of computed tomography, hepatobiliary scintigraphy with iminodiacetic acid, and ultrasonography showed mild, diffuse thickening of the gallbladder wall, mild hepatomegaly and splenomegaly, and pericholecystic fluid. The findings were consistent with a diagnosis of acute or chronic cholecystitis. There was no definite evidence of cholelithiasis, duct dilatation, or abscess (see the figure in the Supplementary Appendix, available with the full text of this article at NEJM.org). The patient received a diagnosis of probable acute cholecystitis. In addition to routine blood and urine cultures, tests were performed for infection with or exposure to the following infectious agents before or during this admission: hepatitis viruses, cytomegalovirus, parvovirus, varicella–zoster virus, HIV, Epstein–Barr virus, Treponema pallidum, ehrlichia, and anaplasma. The tests for all these infectious agents were negative. In the hospital, the patient was initially treated with levofloxacin (750 mg given intravenously every 24 hours) and meropenem (1 g every 8 hours); on July 14, levofloxacin and meropenem were replaced with vancomycin (1.25 g given intravenously every 12 hours), cefepime (2 g given intravenously every 12 hours), and metronidazole (500 mg given intravenously every 6 hours), owing to a concern that worsening thrombocytopenia could have been related to the meropenem. The only tests that had positive results for infection were a blood polymerase-chain-reaction (PCR) assay that was positive for herpes simplex virus (HSV), which was minimally detectable at 300 copies per milliliter, and a tracheal aspirate that was positive for candida.
|
The patient was transferred to our hospital on July 18 for possible liver transplantation, and an adverse-event report related to the gene-therapy trial was filed with the FDA on July 19. Her initial antibiotic treatment was piperacillin with tazobactam (2.25 g given intravenously every 6 hours), but that regimen was switched on July 22 to imipenem (250 mg given intravenously every 6 hours) and daptomycin (180 mg every 48 hours, increased to 450 mg every 48 hours on July 23). The blood vancomycin level was 46.0 µg per milliliter on July 19; on July 22, when it reached 12.7 µg per milliliter, a single dose of 1 g of vancomycin was given intravenously. The patient was jaundiced and had scleral icterus. The initial bladder pressure was elevated, at 22 mm Hg; the central venous pressure was 6 cm of water; and the central venous oxygen saturation was low, at 63%. She had hemodynamic instability requiring transfusion of multiple blood products and continuous venovenous hemofiltration. The retroperitoneal bleeding caused increased abdominal pressure and altered respiratory mechanics (Figure 1A and 1B). Because of the patient's involvement in the gene-therapy trial, the liver-biopsy specimen was evaluated for potential viral changes. There were small areas of necrosis, but no evidence of virus was detected, and there was substantial normal-appearing liver (Figure 2A).
|
|
Results
Blood cultures obtained on July 24 showed positive results for Histoplasma capsulatum on July 26. Fungal stains of the liver-biopsy specimen, which became available after the patient's death, revealed many histoplasma organisms, but without granuloma formation. Results of immunostaining for HSV were negative. Examination of other organs revealed disseminated histoplasma in the lungs, bone marrow, spleen, lymph nodes, kidney, intestine, pancreas, and brain (Figure 2B and 2C). A quantitative enzyme immunoassay for histoplasma antigens, which was performed after the patient's death, showed that a serum sample from May 29 was negative. However, a serum sample from July 2 had a concentration of less than 0.6 ng per milliliter, which was higher than the cutoff value for positivity but lower than the lowest concentration on the calibration curve (MiraVista Diagnostics).7 This finding, along with the symptoms, suggests that the patient had an unrecognized histoplasma infection at the time of the second injection.
The retroperitoneal hemorrhage was extensive; the affected area weighed 3.6 kg, and the hemorrhage caused compression of the lungs. There was segmental infarction of the left kidney. No anatomical source of the bleeding could be identified, and there was no evidence of hemorrhage in other organs. The possibility of an undetected mycotic aneurysm could not be ruled out, since a small aneurysm might not have been evident because of the amount of bleeding. Coagulation tests performed on a blood sample obtained on July 16 were consistent with a consumptive coagulopathy together with some degree of hepatocytic synthetic deficiency (Table 2). The high level of von Willebrand factor could have been due to an acute-phase reaction, although endothelial injury could not be ruled out. Coagulation-factor assays at multiple dilutions did not show an inhibitor. Although the correction phase8 of the dilute Russell's viper–venom time showed a slightly positive result, the hexagonal-phase phospholipid neutralization test was negative, and the levels of anticardiolipin antibodies were not elevated.
|
Results of the PCR analysis for the tgAAC94 vector DNA were consistent with those from preclinical biodistribution studies.9 Results of PCR assays for the AAV rep gene did not show substantial levels of competent virus (i.e., virus that was able to replicate) in the organs. At the injection site, the amount of rep DNA was consistent with the low level of copackaged rep DNA present in the vector product. The serum and plasma protein levels of TNF-
antagonists, as assessed by a radioimmunoassay,10 were not above the expected range for a patient who was receiving doses of systemic adalimumab while also receiving methotrexate (Table 3).
|
Discussion
H. capsulatum is a dimorphic fungus that primarily causes pulmonary disease. For most of her life, this patient lived in a region in which this fungus was endemic. Although clinical manifestations occur in less than 1% of infected persons,12 viable organisms can remain in a latent state, and reactivation can occur when immunity is suppressed,12 such as when a person is being treated with a TNF-
antagonist.13,14,15 The FDA recently emphasized this risk in a warning to health care professionals.16 It is recommended that when TNF-
antagonist therapy is used, patients be screened for tuberculosis, and health care professionals should be vigilant for histoplasmosis in patients who live in areas where the fungus is endemic, although routine screening for histoplasmosis is not recommended, owing to a low incidence of the infection.2,17,18 The case described here is most consistent with reactivation of histoplasmosis, although a recent primary infection cannot be ruled out. The histoplasma antigen can be detected with the use of an immunoassay, with the highest sensitivity found in the urine12,19,20; however, this patient was never tested for urinary histoplasma antigen. Although a specific cause of the retroperitoneal bleeding was not identified, it was probably due to a consumptive coagulopathy from the disseminated histoplasma infection.21,22,23
AAV is a parvovirus and has a single-stranded DNA genome.24,25 The wild-type virus requires a helper virus — either adenovirus or HSV — to replicate, and infection by AAV is not associated with human disease. The wild-type virus has only two genes, rep and cap, both of which were replaced in the investigational agent used in this research study with a promoter sequence and the cDNA sequence equivalent to that for etanercept. The vector is packaged in the AAV2 capsid. The vector is expected to be nonreplicative, even in the presence of a helper virus, and the cDNA is expected to persist for the life of the cell, primarily as an episomal concatemer (i.e., extrachromosomal DNA covalently linked in tandem).26 In this case, analysis of organs for the AAV rep gene did not show substantial levels of competent virus in the organs, so there was no evidence to suggest that there was widespread replication of the AAV virus. The level of vector is consistent with that in preclinical biodistribution studies.9 The assay for total protein TNF-
antagonists showed the expected decline in levels after termination of systemic adalimumab therapy (the patient's last dose of 40 mg was administered within the last 9 days of June). The assay cannot distinguish the TNFR:Fc protein expressed from the intraarticular gene-transfer agent from the anti–TNF-
antibody adalimumab administered systemically; however, for the time points available, the levels were not above those expected with the doses of adalimumab that were administered. Owing to a lack of samples, the vector level in the blood during the 3 weeks after the injection is unknown. However, no vector was found at autopsy 3 weeks after dosing.
Preexisting or induced immunity to the AAV capsid can have a substantial effect on vector levels in the blood.11,27 Eighty percent of the population has neutralizing antibodies against the AAV2 capsid, but usually at a low titer.28 The cellular immune response to AAV vectors warrants further investigation.29,30 In this case, there were no samples available for assessment of the T-cell response.
On the basis of its investigation, the Recombinant DNA Advisory Committee concluded that the patient's death was primarily a result of disseminated histoplasmosis with subsequent bleeding complications and multiorgan failure. The apparent risk factor was the patient's systemic treatment with adalimumab. The contribution of an immune response to the AAV vector could not be evaluated. Although the degree of TNF inhibition from the transgene product could not be determined, the level of TNF antagonist in the bloodstream was not higher than that which would be expected, given the patient's use of systemic adalimumab. Other than the timing of the administration of the gene-therapy agent, none of the available data support a conclusion that the gene-therapy agent contributed to the patient's clinical course or death.
This case highlights issues that are relevant to any clinical trial. When an unexpected illness develops in a study participant, who is the best person to determine whether the illness is related to the treatment under investigation? In this case, the investigator and the sponsor initially determined that the patient's death was unlikely to have been related to the gene-therapy agent but later concluded that it may possibly have been related. A hospital that was not connected with the study reported the illness to the FDA on July 19. Although the investigator on site will probably continue to be the person who decides whether an event is related to the study treatment, it is imperative that samples be collected at the time of admission to the hospital, regardless of the initially perceived relationship, thus permitting follow-up tests to be performed later. The availability of multiple sections of each tissue, frozen and stored individually in separate tubes, may be of value in a retrospective analysis of the cause of death. The recommendations of the Recombinant DNA Advisory Committee included monitoring anti–AAV-capsid T cells before and at multiple times after each injection, developing an assay to specifically quantify the transgene product as distinguished from other biologic agents, screening for signs of active infection, and considering the creation of a trial contact number that can be accessed 24 hours a day.9
The death of this young patient who was receiving multiple forms of TNF inhibitors highlights the risk of opportunistic infections in patients receiving such agents2,16 and the importance of having a well-designed monitoring plan when a patient in a study becomes ill.
Dr. Samulski reports receiving consulting fees from and owning equity and stock options in Ceregene and Asklepios BioPharmaceutical; Dr. Samulski, being named as a coinventor on a number of patent applications for gene delivery, the majority of which are related to adeno-associated virus vectors; and Dr. Mease, receiving consulting fees from Abbott, Amgen, Biogen Idec, Bristol-Myers Squibb, Centocor, Genentech, Roche, Targeted Genetics, Wyeth, and UCB, lecture fees from Abbott, Amgen, Biogen Idec, Bristol-Myers Squibb, Centocor, Genentech, and Wyeth, and grant support from Abbott, Amgen, Biogen Idec, Bristol-Myers Squibb, Centocor, Genentech, Roche, Targeted Genetics, and Wyeth. No other potential conflict of interest relevant to this article was reported.
We thank the patient's family for permission to share this case publicly; Dr. Jay Lozier, Dr. Barry Byrne, Dr. Jacqueline Corrigan-Curay, members of the Recombinant DNA Advisory Committee, and members of the FDA for multiple discussions of the case; Dr. David Paushter for review of radiographic images; and the staff at the sponsoring corporation (Targeted Genetics) for their contributions to the investigation. The sponsor also funded additional specialized testing as part of the autopsy investigation that was performed at reference laboratories under contract.
Source Information
From the Department of Pathology (K.M.F., J.L.M., S.M., J.H.) and the Division of Pulmonary and Critical Care, Department of Medicine (D.K.H.), University of Chicago Medical Center, Chicago; Seattle Rheumatology Associates, Division of Rheumatology Research, Swedish Medical Center, University of Washington School of Medicine, Seattle (P.J.M.); the Gene Therapy Center, University of North Carolina at Chapel Hill, Chapel Hill (R.J.S.); and St. John's Hospital, Springfield, IL (G.A.W.).
Address reprint requests to Dr. Frank at the University of Chicago Medical Center, 5841 S. Maryland Ave., MC 0001, Chicago, IL 60637, or at kfrank{at}uchicago.edu.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Death in a Gene-Therapy Trial
Zaia J. A., Federoff H. J., Hage C. A., Bowyer S., Kleiman M. B.
Extract |
Full Text |
PDF
N Engl J Med 2009;
361:1811-1812, Oct 29, 2009.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2010 Massachusetts Medical Society. All rights reserved. |