1 INTRODUCTION

Transplantation (Tx) in patients with organ failure is a lifesaving procedure that carries intrinsic medical risks. Besides rejection, and development of opportunistic infections and cancers due to immuno-suppression, there is also the possibility of donor-transmitted malignancies. Such incidence ranges between 0.01 and 0.06%— melanoma, colon, and lung cancer being the most common transmitted tumors.^1-3 Donor-derived cancers can either develop early—that is, during the first few weeks after transplantation—or at a later time. In the first instance, tumors are transplanted together with the organ, engrafting notwithstanding HLA mismatch, possibly due to recipient's immunosuppression. Late-onset donor-derived cancers are less well understood and more difficult to recognize and treat.

In July 2013, nephrologists at an Italian hospital reported to the Regional Transplantation Center (RTC) the possible transmission of donor-related Acute Myeloid Leukemia (AML) in the kidney recipients from a common donor. Upon transplantation history evaluation, it was discovered that the liver recipient had already died from AML in November 2012. Hence, we undertook a complete review of donor's and recipients' clinical history, and performed comprehensive molecular analyses to ascertain the AML origin, confirm its transmission from the donor, and tease apart the distinct evolution paths in the recipients. We report here our findings.

2 CASE REPORTS

2.2 Transplant recipients

All transplantations were successful, and no major post-surgical complications occurred. Two recipients (R1, a 57-year-old male, and R2, a 55-year-old female) received the kidneys, a third recipient (R3, a 62-year-old female) received the liver. From the histocompatibility standpoint, the two kidney transplantations were partial HLA-A, B, and DR matches, while the liver recipient was a full mismatch (Table S1). Immunosuppressive therapy for all recipients included—in accordance to best practices—induction with anti-CD25 antibodies before transplantation followed by tacrolimus, mycophenolate mofetil, and prednisone. After diagnosis of AML, immunosuppression was reduced to prednisone.

The two kidney recipients were followed by the same team, while the liver recipient was followed in a different hospital. R1 was diagnosed with AML in May 2013, received treatment, but died of complications in April 2014. R2 was diagnosed with myeloid sarcoma of the transplanted kidney—an extramedullary manifestation of NPM1-mutant AML,⁵ a diagnosis revised to AML upon bone marrow biopsy in June 2013. R2 was treated for leukemia, achieving complete remission until relapse in February 2017. In May 2017, R2 received an HLA-A, B, C, DR, and DQ identical, unrelated, allogeneic bone marrow transplant, but died of treatment-induced complications in December 2017. R3 was diagnosed with AML in September 2012 and died in November 2012 for complications during AML treatment. The clinical follow-up timeline is shown in Figure 1.

3 METHODS

4 RESULTS

4.2 Somatic variants

We performed WES to obtain a detailed characterization of the mutational landscape. Summaries for high-quality variants are reported in Table S3 A–D. Overall, pair-wise correlations based on these variants revealed that post-transplantation samples are highly similar to the donor (R ≅ 0.97 for R1 post-TX and R ≅ 0.81 for R2 post-TX, Figure 2). These findings agree with targeted analyses, confirming the donor origin of AML (Table 1).

Through WES, we also identified the variants shared between the donor and the recipients, as well as those present in some of the samples. Among the 870 shared mutations, 485 were previously reported in COSMIC (Figure 3). This analysis confirmed the NPM1 frameshift mutation identified by targeted sequencing, and revealed additional somatic mutations in genes previously reported as AML drivers,^13-16 including TET2, DNMT3A, RUNX1, CDKN2A, ASLX1, and BCOR among others. Finally, we were also able to confirm the FLT3 ITD mutation in R1 post-transplantation using Pindel¹¹ (Table 1, Figure 4A, and Table S11).

5 DISCUSSION

We have documented for the first time AML transmission from a single donor to three distinct recipients through solid organ transplantation, and its independent clonal evolution, by multi-level evidence encompassing targeted molecular analyses and whole exome sequencing. Although we were unable to detect AML in the donor at donation, collectively, our findings support the notion of a direct engraftment of leukemic cells already present in the donor, rather than the evolution of pre-leukemic cells in the recipients.

The set of AML pathognomonic mutations shared among the donor and the recipients encompassed NPM1, RUNX1, DNMT3A, and TET2, in agreement with their “founding” nature,^13-16 whereas the FLIT3 ITD mutation was only detected in one of the kidney recipients, in agreement with AML clonal selection in this subject. Additional evidence of distinct clonal evolution was provided by the identification of sub-clonal inactivating mutations in other AML drivers (like ASLX1 and BCOR), in important cancer genes (e.g., WT1, ATM, TWIST1, or NOTCH3), and in several genes for which a role in AML still needs to be defined.

Similarly, also copy number analysis results were consistent with an independent AML evolution in the two kidney recipients. While the total CNA burden was low—in agreement with the observed mutational spectrum and consistent with cytogenetically normal AML¹⁷—the assessment of informative allele frequencies revealed a sub-clonal loss of chromosome 6 short arm (encompassing the HLA loci) in one of the recipients. These findings might suggest an additional explanation—in addition to standard immunosuppressants—for tumor cells escaping allorecognition, consistent with findings of loss of the HLA incompatible haplotype in relapsing AML after haploidentical bone marrow transplantation.¹⁸

Our study supports the notion that AML can be transmitted also through solid organ transplantation, and not only through bone marrow and hematopoietic stem cells infusion. Donor-derived acute leukemia and myelodysplastic syndromes are well-documented complications in recipients of bone marrow and peripheral and umbilical cord blood, with an incidence ~0.5%.¹⁹ This complication results from the engraftment of premalignant clones, carrying mutated hematologic cancer genes (like TET2 and DNMT3A),²⁰ that undergo proliferative and self-renewal stress during engraftment. This condition, known as clonal hematopoiesis of indeterminate potential (CHIP),²¹ is a strong risk factor for developing an hematologic cancer, and it is observed in 10% of individuals 65 years or older, while it is rarely seen in young people—hence it is also referred to as age-related clonal hematopoiesis (ARCH).²² Recent studies of CHIP/ARCH mutational landscape have revealed that progression to AML is associated with a higher number of mutations, the involvement of specific genes, and also higher mutation frequencies, suggesting a greater clonal expansion.^{14, 15}

AML occurrence in recipients of solid organ transplantation has been already described. A case of acute promyelocytic leukemia was reported in 1999²³ in a 57-year-old woman 2 years after receiving the liver of a 16-year-old male who died of head injury. The authors demonstrated donor's origin of the AML, but in retrospective analyses found no evidence of leukemia in the donor, concluding that the disease resulted from donor cell transformation after transplantation. Furthermore, this report did not include any information about the recipients of the other organs (if any) and their outcomes. A second case was reported in 2013,²⁴ when a 69-year-old woman died of AML complications 2 years after a double kidney transplantation. Since no signs of leukemia were found in the donor and the liver recipient did not develop AML, also in this case the authors concluded that the disease arose in the recipient. On the contrary, in the cases we have described, the leukemic clones were already present in the donor, as demonstrated by the presence of founder mutations and by their transmission to all three recipients.

Another interesting aspect to consider is the time lapsed before the AML manifested itself in the recipients. Given the extraordinariness of this event, with obvious limitations, we can only refer to what is known about the engraftment of human tumors into immunocompromised mice, in which a lag time of months is not at all infrequent (see Chen et al²⁵). To this end, important aspects in determining engraftment time are the number of injected cells and their viability. We have no information about these factors, but we can assume that the absolute number of leukemic cells within the transplanted organs was low, with some variability—also in terms of viability—between the liver and kidneys, due to the intrinsic differences in volume and microenvironment between these organs. It is worth noting that the liver recipient died earlier than the kidney ones, who were, in turn, diagnosed close in time, suggesting that these latter patients might have received similar numbers of leukemic cells that engrafted in comparable times, notwithstanding the different hosts. Finally, the most relevant difference with xenograft models, is that the recipients were not immunodeficient and that they were only incompletely immunosuppressed. Hence, to fully explain why their immune system did not reject the tumor, we can postulate that additional mechanisms of immune evasion, impossible to ascertain by WES, were also at play (e.g., the expression of immune check-points molecules). Among these, it is interesting to note that tumor cells of one of the recipients lost at least one HLA haplotype.

The incidence of AML following solid organ transplantation is yet to be studied in detail. In a 2014 review,²⁶ Rashidi and colleagues reported 51 AML cases following organ transplantation, of which four, all occurring after liver transplantation, were deemed to be donor-derived. For the additional cases more recently reported, donor origin was not investigated.²⁷ Given the fact that hematopoietic stem cell can survive in transplanted organs,²⁸ the increasing use of older donors, and the prevalence of ARCH later in life, it is not surprising to start observing donor-derived and, with this report, donor-transmitted leukemias also after organ transplantation. The true extent of this complication—ranging from 0.18% to 0.8%²⁶—might be somewhat underestimated, since leukemias in the recipients can also result from immunosuppression, and usually only cases of multiple transmissions from single donors are fully investigated. Notably, this is also true for breast²⁹ and other cancer types,^{2, 29, 30} suggesting the importance of continuous improvements to the reporting guidelines.

For AML and other leukemias, since ARCH/CHIP increases with age, one could speculate that older donors should be considered at high risk. While this is extremely important in bone marrow transplantation, there are no data suggesting that ARCH/CHIP clones will engraft in solid organ transplantation recipients. Hence, despite our findings, we believe that excluding donors based on age is not warranted, since this would result in an unacceptable reduction of the donor pool with a consequent reduction in transplantation survival benefits. Most importantly, under existing guidelines, cancer transmission risk is overestimated, resulting in the exclusion of high-risk donors that could be instead safely used.^{2, 30}

A possible solution could be updating current guidelines to include hematologic tumor screenings for older donors, although systematically performing such testing would be burdensome, since currently—at least in the Italian experience—more than half of deceased donors are already over 65 years of age, suggesting the need for donor risk stratification strategies. Although blood smears do not always allow for cancer detection and bone marrow biopsies aren't usually feasible in emergency settings, our findings strongly suggest that for older potential donors with hematological abnormalities or alterations suggestive of cancer deeper evaluations by an hematologist would be warranted.

It would be different if rapid molecular tests—compatible with donation—were available to exclude cancerous cells presence. While genomic sequencing is feasible for living donors, in transplantation from deceased individuals this is not possible due to the required fast turn-around time. Nevertheless, with current technological advancements, is plausible that including such cancer detection tests in the donor workup is not far in the future. To this end, while such molecular information cannot yet inform donor risk assessment, it could be already useful for implementing enhanced follow-up protocols of the recipients receiving organs from older donors. Finally, it is worth mentioning that predictive cancer risk models based on health records and other clinical parameters—as shown for red blood cell distribution width in AML¹⁵—already exist, potentially enabling a better donor stratification into risk groups without any delay and at no additional cost.

In conclusion, this is the first report and in-depth genomic characterization of AML transmission following solid organ transplantation. We do not have a clear explanation for the time required for the AML to develop in the recipients, except that the number of AML cells transferred into recipient must have been quite small. Our results from whole exome sequencing highlight the potential molecular mechanisms explaining the evolution of the tumor and its escape from immune response in the recipients. Furthermore, all our molecular analyses indicate that leukemic cells were already present in the donor at the time of death, underscoring the need for continuous improvements of donor screening, risk stratification, and recipient follow-up protocols. Since the current donor pool is shifting toward an older age, these efforts are critical to continue minimizing the risk of transmission of cancer and other diseases through organ transplantation.

DISCLOSURE

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.