eTable 1. Standard Antigen Specificities and Broad Specificities
eTable 2. Cross-Reactive Group Antigen Specificities
Customize your JAMA Network experience by selecting one or more topics from the list below.
Gao Y, Twigg AR, Hirose R, et al. Association of HLA Antigen Mismatch With Risk of Developing Skin Cancer After Solid-Organ Transplant. JAMA Dermatol. 2019;155(3):307–314. doi:10.1001/jamadermatol.2018.4983
What is the association between HLA antigen mismatch and the incidence of skin cancer after a solid-organ transplant?
In this secondary analysis of a cohort study of 10 649 organ transplant recipients, HLA antigen mismatch between a donor and a recipient was associated with a protective effect against posttransplant skin cancer after adjusting for age, sex, white race/ethnicity, thoracic organ transplant, year of transplant, and pretransplant history of skin cancer. Higher levels of HLA antigen mismatch reduced the risk of skin cancer after a solid-organ transplant in heart and lung transplant recipients.
Tumor surveillance mechanisms may be activated by HLA antigen mismatch, and well-matched heart and lung transplant recipients may have a higher risk of skin cancer after transplant.
Risk factors for the development of skin cancer after solid-organ transplant can inform clinical care, but data on these risk factors are limited.
To study the association between HLA antigen mismatch and skin cancer incidence after solid-organ transplant.
Design, Setting, and Participants
This retrospective cohort study is a secondary analysis of the multicenter Transplant Skin Cancer Network study of 10 649 adults who underwent a primary solid-organ transplant between January 1, 2003, and December 31, 2003, or between January 1, 2008, and December 31, 2008. These participants were identified through the Scientific Registry of Transplant Recipients standard analysis files, which contain data collected mostly by the Organ Procurement and Transplantation Network. Participants were matched to skin cancer outcomes by medical record review. This study was conducted from August 1, 2016, to July 31, 2017.
Main Outcomes and Measures
The primary outcome was time to diagnosis of posttransplant skin cancer, including squamous cell carcinoma, melanoma, and Merkel cell carcinoma. The HLA antigen mismatch was calculated based on the 2016 Organ Procurement and Transplantation Network guidelines. Risk of skin cancer was analyzed using a multivariate Cox proportional hazards regression model.
In total, 10 649 organ transplant recipients (6776 men [63.6%], with a mean [SD] age of 51  years) contributed 59 923 years of follow-up. For each additional mismatched allele, a 7% to 8% reduction in skin cancer risk was found (adjusted hazard ratio [HR], 0.93; 95% CI, 0.87-0.99; P = .01). Subgroup analysis found the protective effect of HLA antigen mismatch to be statistically significant in lung (adjusted HR, 0.70; 95% CI, 0.56-0.87; P = .001) and heart (adjusted HR, 0.75; 95% CI, 0.60-0.93; P = .008) transplant recipients but not for recipients of liver, kidney, or pancreas. The degree of HLA-DR mismatch, but not HLA-A or HLA-B mismatch, was the most statistically significant for skin cancer risk (adjusted HR, 0.85; 95% CI, 0.74-0.97; P = .01).
Conclusions and Relevance
The HLA antigen mismatch appears to be associated with reductions in the risk of skin cancer after solid-organ transplant among heart and lung transplant recipients; this finding suggests that HLA antigen mismatch activates the tumor surveillance mechanisms that protect against skin cancer in transplant recipients and that skin cancer risk may be higher in patients who received a well-matched organ.
More than 140 000 Americans are solid-organ transplant recipients, and this number continues to increase as organ transplant rates have tripled over the past 25 years.1 Immunosuppressive regimens that reduce the risk of graft rejection increase the risk of cancer, a major source of morbidity and mortality in solid-organ transplant recipients.2,3 The risk of squamous cell carcinoma (SCC) is increased 65-fold and melanoma is increased 3-fold in solid-organ transplant recipients.3-5 Previously defined risk factors for skin cancer after a solid-organ transplant included male sex,6 white race/ethnicity,7,8 receipt of a thoracic organ,4,9 aged 50 years or older at the time of transplant,5,6,10 a long period of immunosuppression,9 past exposure to a high level of UV radiation,11,12 and a pretransplant history of skin cancer.13,14
The HLA antigen major histocompatibility complex (MHC) molecules are expressed on almost all nucleated cells, and the immune system uses them to distinguish the self from the nonself.15 The 3 classical HLA class I MHC molecules are HLA-A, HLA-B, and HLA-C, and the 3 class II MHC molecules are HLA-DR, HLA-DQ, and HLA-DP. These molecules are important in solid-organ transplant because mismatched donor-derived MHC molecules are the stimulus for an alloimmune response resulting in rejection.16 The need for HLA antigen matching in solid-organ transplant varies considerably depending on the type of organ being transplanted and the risk of allograft loss from rejection. The number of HLA antigen mismatches for solid-organ transplant recipients ranges from 0 to 6 and is the sum number of antigen specificities (as defined by serologic nomenclature) mismatched for each of the 2 HLA-A, HLA-B, and HLA-DR alleles.15 It is widely accepted that fewer mismatches result in increased graft survival and fewer rejection episodes.17-24
Molecular mechanisms also play an important role in the defense against the development of cancer. Nonself peptides presented as the HLA-peptide complex of neoplastic cells or cells infected with a virus are recognized by T lymphocytes.25 Altered expression of MHC molecules is a well-described mechanism used by cancer cells to evade immunosurveillance.26 The data on the association between HLA antigen mismatch and skin cancer incidence are limited, and kidney transplant recipients have been the focus.27,28 Here, we evaluate the contribution of HLA antigen mismatch to posttransplant skin cancer risk. We used a population-based data set of skin cancer outcomes combined with national transplant registry data.
This cohort study is a secondary analysis of the previously published Transplant Skin Cancer Network (TSCN) study of posttransplant skin cancer incidence.14 The TSCN is a multicenter initiative that includes 26 transplant centers in the United States with an active collaboration between the dermatology and transplant specialties. This study received approval from the institutional review board of the University of California, San Francisco. Informed consent was not required because the study involved retrospective data with no patient contact.
All adult (≥18 years of age) recipients of a primary transplant at participating transplant centers between January 1, 2003, and December 31, 2003, or between January 1, 2008, and December 31, 2008, were included in the TSCN study. The years 2003 and 2008 were chosen to enable up to 5-year and 10-year follow-up. Recipients of lung, heart, pancreas, liver, and kidney allografts were included. Eligible patients were identified using the Scientific Registry of Transplant Recipients standard analysis files, which are largely based on the data collected by the Organ Procurement and Transplantation Network (OPTN) as of December 2013. This secondary analysis was conducted from August 1, 2016, to July 31, 2017.
The primary outcome was time to diagnosis of posttransplant skin cancer, including SCC, melanoma, and Merkel cell carcinoma. Skin cancer diagnoses and dates were captured through a comprehensive primary medical record review, as detailed in the TSCN study.14
The HLA antigen mismatch can be described in more than 1 way. Standard HLA antigen typing for a transplant is reported at the antigen level (eg, HLA-A2).29 The OPTN database contains HLA antigen data for recipients and donors, including HLA-A, HLA-B, and HLA-DR antigens as well as the number of mismatches for this study cohort based on laboratory-specific typing methods in the United States. Major histocompatibility complex molecules were named sequentially as they were identified by serologic methods, but a change from serologic typing to more specific typing resulted in an evolution of the nomenclature, as the number of MHC molecules has rapidly increased.30 Identification of new MHC molecules has split the existing antigens, often distinguishing allele groups that make up a family of related antigens. The HLA-B40, for example, was split into 2 groups and renamed HLA-B60 and HLA-B61. The HLA-B40 is, therefore, the parent antigen of HLA-B60 and HLA-B61.31 There is also an overlapping nature in the way the immune system recognizes MHC molecules. Immunization to a single HLA alloantigen can result in antibodies that recognize other MHC molecules that contain similar sequence motifs (public epitopes).32 These class I antigens, which belong to the same cross-reactive group (CREG), contain shared amino acids that correspond to regions relevant to the antibody-binding sites. Several antigens contain more than 1 public determinant and belong to more than 1 CREG.33,34 We calculated HLA antigen mismatch levels using 3 methods: standard antigen mismatch, broad specificity mismatch, and CREG mismatch.
Standard antigen mismatch was calculated according to section 4.10 of the 2016 OPTN Policies.35 A mismatch was defined as an antigen present in the recipient but not in the donor or an antigen present in the donor but not in the recipient. We recoded these mismatch levels from the raw HLA antigen types listed in the OPTN database, as the current typing methods have changed the organization and naming of HLA antigens and created more stringent matching criteria than the guidelines used for the coding of the cohort at the time of transplant in 2003 or 2008 (eTable 1 in the Supplement). Standard antigen mismatch levels range from 0 (perfect match, or 6 out of 6 matches) to 6 (0 out of 6 matches) for HLA-A, HLA-B, and HLA-DR.
Splits were recoded to reflect the broad parent group.31 For example, HLA-A29, HLA-A30, HLA-A31, HLA-A32, HLA-A33, and HLA-A74 are splits of the HLA-A19 parent antigen and were all recoded as HLA-A19 (eTable 1 in the Supplement). The sum of the mismatches for the data recoded according to broad specificities determined the overall broad specificity mismatch level, ranging from 0 to 6. Broad specificity grouping captures more antigens that are functionally matched to the recipient; thus, the mismatch frequency is lower.
In this analysis, a CREG mismatch was defined as the presence of any CREG in the donor but not in the recipient (eTable 2 in the Supplement).36,37 The CREG coding only applies to class I antigens (HLA-A and HLA-B), and the corresponding CREG mismatch level ranges from 0 to 7. Mismatches for HLA-A and HLA-B were combined to establish an overall CREG mismatch level. Furthermore, CREGs increase the sensitivity of detecting functionally matched donor and recipient MHC molecules, reducing mismatch frequency compared with standard and broad specificity matching.
Multivariate Cox proportional hazards regression was used to test the associations between the degree of HLA antigen mismatch calculated with all 3 coding methods (standard antigen mismatch, broad specificity mismatch, and CREG mismatch) and time to development of skin cancer after transplant. The degree of HLA antigen mismatch was included in the analysis as a linear variable to determine the association between HLA-A, HLA-B, and HLA-DR mismatches and skin cancer risk. A subgroup analysis was performed to examine the association of HLA antigen mismatch with skin cancer risk by transplanted organ type. Hazard ratios (HRs) and 95% CIs were adjusted for known risk factors associated with skin cancer development during the posttransplant period, including age 50 years or older, male sex, thoracic organ transplant, year of transplant, pretransplant history of skin cancer, and white race/ethnicity, as reported in the TSCN study.14 The proportional hazards assumption was tested using the Schoenfeld test. The HRs were calculated in the presence of competing risk, defined as death from all causes.38 Competing risk regression yielded similar HRs as those obtained with the Cox proportional hazards regression.
Sensitivity analyses were conducted to assess the implication of missing predictor variables: skin cancer before transplant (39%), standard mismatch level (32%), broad specificity mismatch level (32%), and CREG mismatch level (27%). Through inverse probability weighting, a logistic regression model for missing predictor variables was developed, and the analysis was adjusted for missing data using the weighting scheme.39 To acknowledge the clustering effects by transplant center, we applied the Huber-White method of cluster-robust SE to the Cox proportional hazards regression analysis.40 A level of 5% was used to determine statistical significance. Statistical analyses were weighted and performed using STATA, version 14.2 (StataCorp LLC).
A total of 10 649 organ transplant recipients (3872 women [36.4%] and 6776 men [63.6%], with a mean [SD] age of 51  years) contributed 59 923 years of follow-up. Details of all demographic characteristics of the TSCN study cohort and risk factors for skin cancer are published elsewhere.14Table 1 summarizes the mean and SD for degree of HLA antigen mismatch as calculated through the 3 coding methods for cohort demographic variables. As expected, increasing the sensitivity to detect functional matches by using broad specificities and CREGs reduced the number of mismatches between donors and recipients. The overall mean (SD) HLA antigen mismatch was 4.3 (1.6) with the standard antigen mismatch method, 4.1 (1.5) with the broad specificity mismatch method, and 2.0 (1.3) with the CREG mismatch method. Thoracic organ transplant recipients had a higher mean (SD) HLA antigen mismatch level compared with abdominal organ transplant recipients, using the standard antigen mismatch method (4.9 [1.0] vs 4.1 [1.6]) and the broad specificity mismatch method (4.7 [1.0] vs 3.9 [1.6]). Table 2 demonstrates how the 3 coding methods affected the frequency distribution of HLA-A, HLA-B, HLA-DR, and combined HLA antigen mismatch levels.
A great number of mismatched alleles, as identified through the standard antigen mismatch and broad specificity mismatch methods, was associated with a protective effect against the development of skin cancer after transplant (Table 3). For each additional mismatched allele, a 7% to 8% reduction in skin cancer risk was found through standard antigen mismatch (adjusted HR, 0.93; 95% CI, 0.87-0.99; P = .01) and broad specificity mismatch (adjusted HR, 0.92; 95% CI, 0.86-0.98; P = .008) methods. No association with skin cancer risk was found through the CREG mismatch method (adjusted HR, 1.01; 95% CI, 0.94-1.09; P = .71). The adjusted HRs for SCC and melanoma were similar to those for any skin cancer (Table 3), but the HR was not statistically significant for melanoma because of the smaller number of events. The adjusted HR for developing any skin cancer was statistically significant for HLA-DR mismatch (adjusted HR, 0.85; 95% CI, 0.74-0.97; P = .01) but not for HLA-A or HLA-B mismatch.
A priori subgroup analysis of the association between HLA antigen mismatch and skin cancer risk by organ type determined that the protective effect of standard antigen mismatch was seen among lung (adjusted HR, 0.70; 95% CI, 0.56-0.87; P = .001) and heart (adjusted HR, 0.75; 95% CI, 0.60-0.93; P = .008) transplant recipients. We did not detect this association in recipients of liver, kidney, or pancreas transplant. Lung transplant recipients had the highest incidence of posttransplant skin cancer (160 of 545 (29.4%); Table 4). Among kidney transplant recipients, the overall prevalence of skin cancer was higher among those with 0 HLA antigen mismatch compared with those with 1 or more HLA antigen mismatch (11.0%), but this outcome was not statistically significant (14.5% vs 11.0%; P = .13).
Recipients of solid-organ transplant require lifelong immunosuppression to prevent rejection, but immunosuppression may increase the risk of skin cancer.3 Tumor surveillance is a biologic system through which the immune system can specifically identify and eliminate tumor cells that express tumor-specific antigens. Through a process called heterologous immunity, T cells are known to exhibit cross-reactivity41: T cells that recognize allogenic MHC molecules cross-react with vesicular stomatitis virus,42 influenza virus,43 and Epstein-Barr virus,44 among others. The same mechanism may allow T cells to cross-react with antigens expressed by malignant cells.
This study is a national population-based assessment of the association between the HLA antigen mismatch between donors and recipients and the risk of skin cancer after solid-organ transplant. Our results indicated that a 7% to 8% decrease in the risk of skin cancer occurred after solid-organ transplant for each additional HLA allele mismatch, including standard antigen and broad specificity mismatches, and this decrease was seen primarily in recipients of heart and lung transplant. These results suggest that chronic exposure to mismatched alloantigens may stimulate tumor surveillance mechanisms in solid-organ transplant recipients that protect against the development of skin cancer. The similar protective effect observed between the standard and broad specificity mismatches suggests that the degree of allelic molecular divergence does not correlate with the extent of tumor surveillance in the recipient.
In contrast, a CREG mismatch was not statistically significantly associated with skin cancer. Major histocompatibility complex molecules that belong to the same CREG carry both public and private epitopes, indicating that antibodies against certain conserved amino acid residues on class I antigens may not be involved in the immune mechanisms that target cross-reactive skin cancer antigens. In the setting of public epitope mismatch, antibodies in the recipient target regions of donor antigens that are found primarily exterior to the peptide binding groove of class I MHC molecules.32 Immunogenic private epitopes are presented to specific alloreactive T cells with subsequent cytokine secretion and humoral activation, whereas public epitopes may be less immunogenic and even contribute to the development of tolerance rather than reactivity.45
The importance of HLA antigen matching for organ transplant has generally been deemphasized because of the effectiveness of immunosuppressive drugs.46 Previous studies have separately shown that greater HLA antigen mismatch necessitates a higher level of pharmacologic immunosuppression47,48 and that greater immunosuppression is associated with higher skin cancer risk.49 However, the exact association between the degree of HLA antigen mismatch and the level of immunosuppressive medications has not been explicitly established and varies greatly on an individual basis.
Our subgroup analysis shows that an increased degree of HLA antigen mismatch has a protective effect against skin cancer in heart and lung transplant recipients. The reason for this protection may be that the allocation and transplant systems put less emphasis on HLA antigen matching for lung and heart allografts than for other organs.50 In general, heart and lung allografts evoke a stronger immune response,51 and heart and lung transplant recipients receive a higher burden of immunosuppression to prevent rejection.51 In our cohort, lung transplant recipients had the highest incidence of skin cancer, followed by heart transplant recipients. However, these results show that receiving an allograft with a higher degree of HLA antigen mismatch reduces the risk of skin cancer in heart and lung transplant recipients, the transplant population that is the most immunosuppressed. In the setting of a highly immunogenic transplanted organ, elevated levels of immunosuppression may increase the overall risk of skin cancer, but potential tumor surveillance mechanisms that remain intact despite immunosuppression may be more active in protecting heart and lung recipients with a higher level of HLA antigen mismatch.
The association between a lower degree of HLA antigen mismatch, particularly HLA-DR mismatch,52,53 and graft survival is well established.17-24 Results of this study show that HLA-DR mismatch has the strongest protective effect against developing skin cancer. The HLA-DR may be a more potent inducer of CD4+ T cells, which enhance the alloimmune response.54 In addition, the role that HLA-DR mismatch plays in reducing skin cancer risk suggests that antigen presentation is important in skin cancer tumor surveillance. Some studies have shown that human papillomavirus (HPV) is associated with cutaneous SCC, especially in immunosuppressed individuals.55,56 HLA-DR has been shown to be involved in the antigen presentation of HPV peptides to tumor-infiltrating lymphocytes.57 HLA-DR mismatch is likely associated with the activation of antigen-presenting cells that help control HPV-induced tumorigenesis of SCC in transplant recipients. The process of heterologous immunity in the setting of T-cell receptor degeneracy may also play a role: T-cell receptors bind and react to allogenic MHC molecules, which cross-react with HPV antigen–presenting cells in the skin of transplant recipients.
The main limitations of this cohort study were the retrospective nature of the database analysis and the incomplete data on skin cancer incidence and typing of other HLA antigen loci. We addressed this limitation through inverse probability weighting to account for missing data. The OPTN registry does not contain reliable data on rejection episodes or immunosuppressive treatment regimens; thus, we adjusted for thoracic (rather than abdominal) transplant in the model as a proxy for the higher levels of immunosuppression in heart and lung transplant recipients. Last, selection bias may have been introduced by participants who were from academic transplant centers. Compared with all transplant patients in the United States, a larger proportion of participants were older, white, and recipients of a thoracic transplant.14 These covariates were all included in the adjusted regression models.
This study examined the association between HLA antigen mismatch and skin cancer risk in a population cohort of solid-organ transplant recipients. The results show that increasing the degree of HLA antigen mismatch appears to reduce the risk of skin cancer in solid-organ transplant recipients by 7% to 8% for each mismatched allele, and this association was seen among heart and lung transplant recipients. The HLA-DR mismatch seemed to play the strongest role. Internists and dermatologists who treat organ transplant recipients should be aware that the risk of skin cancer may be higher in patients who underwent thoracic transplant and received a well-matched organ. Follow-up studies should focus on determining how the burden of immunosuppression affects the association between the degree of HLA antigen mismatch and skin cancer risk as well as the molecular mechanisms that explain the protective effect of HLA antigen mismatch.
Accepted for Publication: November 8, 2018.
Corresponding Author: Sarah T. Arron, MD, PhD, Department of Dermatology, University of California, San Francisco, 1701 Divisadero St, Box 316, San Francisco, CA 94143-0316 (firstname.lastname@example.org).
Published Online: January 23, 2019. doi:10.1001/jamadermatol.2018.4983
Author Contributions: Dr Arron had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Gao, Nowacki, Maytin, Vidimos, Arron.
Acquisition, analysis, or interpretation of data: Gao, Twigg, Hirose, Roll, Nowacki, Maytin, Rajalingam, Arron.
Drafting of the manuscript: Gao, Hirose, Roll.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Gao, Nowacki, Arron.
Obtained funding: Gao, Arron.
Administrative, technical, or material support: Gao, Hirose, Vidimos, Arron.
Supervision: Twigg, Hirose, Roll, Vidimos, Arron.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported in part by the Melanoma Research Foundation. The Transplant Skin Cancer Network was funded by the American Academy of Dermatology and by Galderma.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Data for this study were obtained from the United Network for Organ Sharing Organ Procurement and Transplantation Network, which is supported in part by contract 234-2005-370011C from the Health Resources and Services Administration.
Additional Information: We would like to extend our thanks to all of the Transplant Skin Cancer Network investigators: Giorgia L. Garrett, MD, University of California, San Francisco Medical Center; Oscar Colegio, MD, PhD, Yale New Haven Hospital; Clara Curiel, MD, University of Arizona Medical Center; John R. Griffin, MD, Baylor All Saints and Baylor University Medical Center; Conway C. Huang, MD, University of Alabama Hospital; Anokhi Jambusaria, MD, Mayo Clinic Florida; S. Brian Jiang, MD, University of California, San Diego Medical Center; Justin J. Leitenberger, MD, Oregon Health and Science University; Rajiv I. Nijhawan, MD, Parkland Memorial Hospital and Parkland Southwestern Hospital; Shari Ochoa, MD, MS, Mayo Clinic Arizona; Edit B. Olasz, MD, PhD, Medical College of Wisconsin; Clark Otley, MD, Mayo Clinic Rochester; Arisa Elena Ortiz, MD, University of California, San Diego Medical Center; Vishal Anil Patel, MD, FAAD, FACMS, New York Presbyterian/Columbia University Medical Center; Melissa Pugliano-Mauro, MD, University of Pittsburgh; Chrysalyne D. Schmults, MD, MSCE, Brigham and Women's Hospital; Sarah E. Schram, MD, University of Minnesota Medical Center; Thuzar Shin, MD, PhD, University of Pennsylvania; Seaver Soon, MD, Scripps Green Hospital; Teresa Soriano, MD, University of California Los Angeles Medical Center; Divya Srivastava, MD, Parkland Memorial Hospital and Parkland Southwestern Hospital; Jennifer Stein, MD, New York University Medical Center; Kara Sternhell-Blackwell, MD, Barnes-Jewish Hospital; Stan Taylor, MD, Parkland Memorial Hospital and Parkland Southwestern Hospital; Allison T. Vidimos, MD, Cleveland Clinic Foundation; Peggy Wu, MD, Beth Israel Deaconess Medical Center; and Sarah T. Arron, MD, PhD, University of California, San Francisco Medical Center.
Disclaimer: The views expressed herein are those of the authors and do not reflect the official policy or position of the US Department of Health and Human Services, nor does the mention of trade names, commercial products, or organizations imply endorsement by the US government.
Create a personal account or sign in to: