Post-bone marrow transplant thrombotic microangiopathy

Abstract

Thrombotic microangiopathy (TMA) is a systemic disease characterized by microangiopathic hemolytic anemia, thrombocytopenia and organ failure. Post-bone marrow transplant TMA (post-BMT TMA) is a life-threatening condition that has been reported to afflict between 0.5 and 63.6% of BMT patients. The incidence of post-BMT TMA is affected by evolving therapies such as conditioning regimens. The etiology of post-BMT TMA is thought to be multifactorial, including the effects of immunosuppressive agents, viral infections, TBI and GvHD. A growing body of evidence highlights the importance of complement system activation and endothelial damage in post-BMT TMA. Although plasmapheresis has commonly been used, its therapeutic rationale for the majority of post-BMT TMA cases is unclear in the absence of circulatory inhibitors. It has become possible to target complement activation with eculizumab, a drug that blocks the terminal complement pathway. Early studies have highlighted the importance of anticomplement therapies in treating post-BMT TMA. Moreover, finding complement gene mutations may identify patients at risk, but whether such patients benefit from prophylactic anti-complement therapies before BMT remains to be studied. This review focuses on diagnostic criteria, pathophysiology, treatment and renal outcomes of post-BMT TMA.

INTRODUCTION

Thrombotic microangiopathy (TMA) is a well-described, potentially lethal complication seen in patients undergoing bone marrow transplant (BMT).¹ Post-BMT TMA is characterized by microangiopathic hemolytic anemia, thrombocytopenia and multi-organ failure.² The term TMA defines a lesion of thickening of arteriole and capillary walls with prominent endothelial damage, subendothelial accumulation of proteins and cell debris, and obstruction of vessel lumina by fibrin and platelet-rich thrombi.³ A number of pathogenic factors have been implicated in post-BMT TMA, including immunosuppressive agents such as CNIs^4–7 and OKT3,⁸ as well as systemic viral infections, including cytomegalovirus,⁹ parvovirus,¹⁰ adenovirus and influenza A virus.¹¹ TBI¹² and GvHD¹³ may also have a role in the pathogenesis of post-BMT TMA.

Several diseases comprise the category of TMA, including thrombotic thrombocytopenic purpura (TTP), typical (Shiga toxin-associated) HUS and atypical HUS (aHUS). However, the pathogenesis of post-transplant TMA reflects that of aHUS.³

INCIDENCE OF POST-BMT TMA

Although the association between BMT and TMA is well documented,^1,2,14–16 the incidence of post-BMT TMA ranges between 0.5 and 63.6%.¹⁷

In a retrospective study, Iacopino et al.¹⁸ evaluated via questionnaire 4334 patients who underwent BMT between 1985 and 1995. They identified nine cases of TMA (0.51%). Rabinowe et al.² diagnosed TMA in 16 (9.5%) out of 168 BMT recipients who met the definition in their study. Twenty-two patients (5%) were diagnosed with TMA by Fuge et al.¹⁹ among 436 patients who underwent allogeneic BMT.

In a more recent study, Shayani et al.²⁰ published a case series of 177 patients who underwent allogeneic hematopoietic stem cell transplantation at City of Hope. Out of the 177 patients studied, 30 (17%) were diagnosed with definite TMA, and an additional nine patients (5%) were classified as probable TMA based on the institutional diagnostic criteria.

The wide range of reported incidence rates for post-BMT TMA is related to the inability to obtain a tissue biopsy and therefore a reliance on clinical parameters. Furthermore, as conditioning regimens have evolved over time, they have had an effect on the incidence of post-BMT TMA. Therefore, new studies are required to elucidate the nature of this disease better in the current area of BMT.

CLINICAL CRITERIA FOR POST-BMT TMA

Several societies have attempted to create criteria for diagnosing post-BMT TMA (Table 1). According to the International Working Group,²¹ five criteria must be met for the diagnosis: schistocytes’ comprising of >4% of the erythrocyte population in a peripheral blood smear; prolonged or progressive thrombocytopenia; an increase in the serum lactate dehydrogenase; a decrease in the serum haptoglobin level; and increased transfusion requirements. The Blood and Marrow Transplant Clinical Trials Network²² requires four criteria for the diagnosis: more than two schistocytes per high-power field; an increase in the serum lactate dehydrogenase level; a doubling of the serum creatinine (SCr) level from baseline; and a negative direct Coombs test. A third set of criteria created by City of Hope²⁰ includes the presence of schistocytes and persistent nucleated RBCs, prolonged or progressive thrombocytopenia, a serum lactate dehydrogenase level greater than twice the upper limit of normal, a decrease in the serum haptoglobin and >50% increase of SCr above baseline. The fourth set of criteria put forward by Overall Thrombotic Microangiopathy Grouping²³ is similar to the International Working Group criteria, except that it does not include transfusion status and instead requires a negative direct Coombs test. The kidney is one of the most critical organs affected by this disease, emphasizing the importance of monitoring renal function via SCr or estimated glomerular filtration rate. Jodele et al.²⁴ showed that in the early post-transplant phase, screening for proteinuria and monitoring urine protein/creatinine ratios can offer diagnostic and prognostic value for patients who are likely to develop TMA after transplantation. This study demonstrated the importance of close monitoring of urinary markers of kidney function, in addition to serum markers.

Table 1.

Definition of TMA following BMT: comparison of four different sets of criteria

Parameter	LeukemiaNet International Working Group	Blood and Marrow Transplant Clinical Trials Network	City of Hope (COH)	Overall Thrombotic Microangiopathy (O-TMA) Grouping
LDH	Increased	Increased	>2× upper limit of normal	Increased
Platelet Count	<50 ×109/L or <50% of normal baseline	-	<50 ×109/L or ≥ 50% decrease from previous counts	<50 ×109/L or <50% of normal baseline
Schistocytes	>4%	>2 per high-power field	Presence of schistocytes, persistent nucleated RBC’s	>2 per high-power field
Creatinine	-	2 × baseline	>1.5 × baseline	-
Haptoglobin	Decreased	-	-	Decreased
Transfusion	Increased	-	-	-
Direct Coombs Test	-	Negative	-	Negative

Abbreviations: BMT = bone marrow transplant; LDH = lactate dehydrogenase.

Since serum haptoglobin can be elevated in many post-transplant inflammatory conditions, its specificity as a marker may be reduced.

Given that the pathogenesis of post-BMT TMA reflects that of aHUS³, its diagnosis requires the exclusion of Shiga toxin-producing Escherichia coli infection and the absence of reduced ADAMTS13 activity.

It is important to note that TMA could be confined to the organ (that is, transplanted kidney) without having systemic manifestations. Schwimmer et al.²⁵ reviewed 1219 biopsy reports of 742 kidney and kidney–pancreas transplants performed during 15 years and found 21 biopsy-confirmed cases of TMA. In total, 62% of these patients had systemic TMA with hemolysis and thrombocytopenia, whereas 38% had TMA localized only to the graft. Tissue characterization would be critical to improve the diagnostic yield of TMA. However, the majority of these patients are at high risk for bleeding complications following biopsy of the kidney. In a recent study, Magro et al.²⁶ used skin biopsies to diagnose aHUS, and they found extensive microvascular depositions of C5b-9, which supported the diagnosis.

Finally, the importance of routine genetic screening for complement gene mutations should be examined in future studies.

PATHOPHYSIOLOGY OF POST-BMT TMA

Post-BMT TMA is a systemic disease which affects virtually all organs. To understand the pathogenesis of this multi-organ disease, it is important to discuss the localization and compartmentalization of complements.

COMPLEMENT SYSTEM

The complement proteins comprise a part of the innate immune system, which initiates and mediates inflammatory reactions, participates in the clearance of endogenous waste products and labels invading microbes for removal by phagocytes.²⁷

KINETICS OF COMPLEMENT SYSTEM AND ITS IMPLICATION IN MULTI-ORGAN INVOLVEMENT

Complement is a system of more than 30 proteins found in plasma and on cell surfaces. These proteins amount to more than 3 g /L and constitute more than 15% of the globular fraction of plasma.²⁸ Components of the classical, alternative and terminal pathways of complements are all present in plasma at concentrations that range from as low as 2 μg/mL (factor D) to as high as 2 mg/ml (C3).²⁹ Complement components have been shown to be cleared by urinary excretion in many studies via detection in urine samples.^30,31

Even though the liver is known to be the main source of complements,^32,33 extrahepatic complement synthesis has been demonstrated in several studies.^34–36 Brauer et al.³⁷ identified the liver as a primary, but not the only source of C6 in the rat. Complement factors have been shown to be secreted by human umbilical vein endothelial cells,³⁸ astrocytes,^39,40 glomerular and renal tubular epithelial cells, mesangial cells and renal vascular endothelial cells,^41–43 as well as the placenta.^44,45 The primary source of plasma factor D is the adipocyte.⁴⁶ C7 is not primarily derived from hepatocytes⁴⁷ but produced mainly by monocytes and tissue macrophages.⁴⁸ Local complement synthesis provides a rapid defense against microbial invasion and these proteins evoke a response comprised of cells and soluble mediators of inflammation.^34,49 These data may explain the involvement of virtually all organs by TMA, arguing in favor of systemic anticomplement strategies, rather than localized therapies (for example, small interfering RNA).

COMPLEMENT ACTIVATION IN TMA

The complement system is activated by three pathways: the classical pathway; the alternative pathway; and the lectin pathway.²⁷ The classical complement pathway typically requires antigen/antibody complexes (that is, IgG or IgM) for activation, whereas the mannose-binding lectin pathway can be activated by microbes with terminal mannose groups, and the alternative pathway is activated by antigens in the absence of antibody, such as those produced by bacteria, fungi, viruses and tumor cells (Figure 1).

Figure 1.

Schematic representation of complement system.

Spontaneous hydrolysis of an internal thioester bond in native C3 initiates the alternative complement pathway.^50–52 It was estimated by Sahu and Lambris⁵² that ~ 0.5% of total C3 exists at any given instant in this hydrolyzed form. The rate of C3 (H2O) formation has been measured between 0.2 and 0.4% per hour.⁵⁰

Meehan et al.⁵³ stained 1101 consecutive renal allograft biopsies for C4d over a period of 51 months. Only six out of 182 C4d+ biopsies were reported as TMA-positive. However, recently Chua et al.⁵⁴ examined 42 renal sections with histologically confirmed TMA. They found that C4d deposits were present in 88.1% of these tissues. They also observed classical pathway activation in 90.5% of the TMA cases. These findings suggest that the alternative pathway should not be the sole pathway investigated for possible diagnostic biomarkers of TMA, as C4d is a marker of the classical pathway.

Tissue characterization of TMA in non-transplant patients who can tolerate organ biopsy can provide key information on the relative importance of these pathways.

ROLE OF GENETIC MUTATIONS IN POST-BMT TMA

The pathogenesis of post-BMT TMA resembles aHUS, which is thought to be mediated by an overactivation of the alternative pathway of complement.^55–58 Complement system mutations can be implicated in up to 52% of patients who have aHUS.⁵⁹ Many different variants in complement genes have been reported to predispose to aHUS,^60–63 such as complement factor H (CFH),^64,65 factor I (CFI),^66,67 C3,⁶⁸ thrombomodulin (THBD),⁶⁰ membrane cofactor protein (MCP or CD46)^69,70 and factor B (CFB).⁷¹ It was shown that patients who have a mutation in MCP carry the best prognosis, as compared with other genetic abnormalities, and almost all of the above genetic mutations are heterozygous.⁶⁰ Mutations in CFH are the most frequent genetic abnormalities in aHUS patients, as they are responsible for 20–30% of cases.^60–63 Patients with factor H mutations have the worst outcomes in comparison to those with other mutations.⁶⁰ Noris et al.⁶⁰ screened CFH autoantibodies in 149 patients with aHUS and detected eight patients without mutations and two with CFH mutations in 10 sporadic cases.

Jodele et al.⁷² performed a hypothesis-driven analysis of 17 different genes that participate in complement activation, during a prospective study they conducted of patients who developed TMA following BMT. They used gene expression profiling to examine the functional significance of variants in these genes. Among 77 patients who underwent genetic testing, 34 had TMA. 65% of the patients with TMA had at least one gene variant, as compared with 9% of the patients without TMA.

Studies to understand better the importance of identifying genetic mutations before BMT remain to be done. Patients with these mutations may require lifetime therapies with complement inhibitors, especially in the presence of inciting factors (that is, immunosuppressive drugs). In addition, identifying BMT patients who are at risk may lead to the prophylactic use of complement inhibitors in these patients.

PATHOGENIC COMPONENTS OF POST-BMT TMA

The endothelium is a highly active tissue that is responsible for regulating vascular tone, coagulation and inflammation. Disruption of endothelial function can cause TMA.⁷³

The calcineurin inhibitors (CNI), such as cyclosporine and tacrolimus, are associated with TMA in both transplant and non-transplant patients.⁷⁴ CNIs cause direct damage to endothelial cells by decreasing the production of prostacyclin and nitric oxide, increasing thromboxane A2, and reducing the formation of activated protein C.^75,76 Renner et al.⁷⁷ showed that endothelial cells exposed to cyclosporine, in vitro and in vivo, release microparticles that activate the alternative pathway of complement. They also found that factor H did not bind to the surface of these microparticles released from cyclosporine-treated endothelial cells, demonstrating a decreased factor H activity. To investigate the effect of tacrolimus, they obtained plasma from renal transplant patients before transplant as well as 2 weeks after starting treatment with tacrolimus, and the number of endothelial microparticles was higher in all patients following treatment with tacrolimus. Finally, they found that alternative pathway-deficient mice were protected from cyclosporine-induced renal and vascular injury.

Kim et al.⁷⁸ demonstrated that cyclosporine-induced renal injury increased the expressions of C3, C4d and MAC (C9), and these were accompanied by increases in complement regulatory proteins (CD46 and CD55), which may be a compensatory response against activated complements.

Activated complement factors can also disrupt endothelial cells by exposing them to granulocytes, which can lead to the release of oxygen radicals by the granulocytes. C5a binds to receptors on endothelial cells and neutrophils (PMNs), inducing P-selectin and activating integrin CD11b/CD18, followed by the release of reactive oxygen species, which mediate endothelial injury.⁷⁹ Sacks et al.⁸⁰ incubated ⁵¹Cr-labeled endothelial cell monolayers with a combination of PMNs and activated serum complement, and they observed that these cells released a higher amount of ⁵¹Cr than cells incubated in the absence of PMNs and complement. They also observed that the absence of either the PMNs or the activated complements prevented endothelial damage.

Several infective agents including cytomegalovirus, adenovirus, Aspergillus, parvovirus and influenza A virus have been associated with post-BMT TMA by causing endothelial damage. In particular, adenovirus expresses a fms-like tyrosine kinase that binds VEGF, leading to TMA.⁸¹

ROLE OF TBI IN THE DEVELOPMENT OF POST-BMT TMA

TBI can cause vascular endothelial injury, and it was shown to be a risk factor for TMA following transplant in many studies.⁸² In both TMA and non-TMA patients, this therapy can cause a loss of kidney function over time. Lawton et al.¹² showed that TBI after BMT places patients at higher risk for developing renal dysfunction. Eighteen months following BMT, 26% of 72 patients who were treated with TBI without renal shielding developed late renal dysfunction as compared with only 6% of 71 patients who were treated with TBI with renal shielding.

ASSOCIATION BETWEEN GVHD AND POST-TRANSPLANT TMA

Changsirikulchai et al.¹³ showed that transplant-associated TMA was four times higher in patients with acute GvHD than in patients without it. It was concluded that endothelial injury may result from circulating cytokines, coagulation pathway activation, low level of VEGF or direct endothelial damage from cytotoxic donor T cells. The relationship between GvHD and the complement system is still unclear.

RENAL OUTCOME OF POST-BMT TMA

TMA is a systemic disease and mostly causes cardiac and neurological problems. However, renal involvement is a dominant feature of post-BMT TMA.⁶⁰ Patients who experience TMA following BMT can have various complications such as hypertension, proteinuria, congestive heart failure, peripheral edema and more serious complications such as ESRD or related death.^2,59,60 Long-term renal complications of post-BMT TMA can lead to significant heart disease, which is a major cause of morbidity and mortality in BMT patients.⁸³

Sellier-Leclerc et al.⁵⁹ observed death or ESRD within 1 year of onset in 17 (37%) of the 46 TMA patients they studied. Noris et al.⁶⁰ found that 81 (29.6%) out of 273 patients with TMA developed ESRD, while 21 (7.6%) patients died following the first episode of TMA. The risk for a poor outcome was found to be related significantly to SCr level at the first episode.⁵⁹

Despite some improvements in overall outcomes, renal dysfunction after BMT remains a significant complication, and regardless of the cause, up to 8% of BMT patients require renal replacement therapy.^84,85

Schwarz et al.⁸⁶ reported the results of 105 renal biopsies from 101 non-renal transplant recipients including liver, lung, bone marrow and heart. Patients with transplant-associated TMA, regardless of transplant type, had the shortest kidney survival.

Glezerman et al.⁸⁷ conducted a retrospective analysis of 100 adult patients who underwent allogeneic BMT. Compared with patients without transplant-associated TMA, those diagnosed with TMA (by biopsy or clinical criteria) were 4.3 times more likely to develop CKD.

TREATMENT OF POST-BMT TMA

Primary therapy for post-BMT TMA involves withdrawal of the offending agent. In a study by Kim et al.⁸⁸ in 2015, 50–63% of patients with post-BMT TMA responded to withdrawal of CNIs. Although plasma exchange (PE) has been used in severe BMT cases, it does not appear to be as effective in the management of post-BMT TMA,⁸⁹ in contrast to its effectiveness in classical TTP. PE is commonly used as an initial therapy for patients who present with TMA, not because it is more effective in patients with aHUS, but because it is difficult to differentiate between TTP and aHUS accurately at the time of presentation. Even though it represents a potentially curative treatment, creating a sense of urgency to initiate the treatment, PE carries a high risk of serious complications.⁹⁰ PE had been used as a main treatment for years because of its ability to remove mutated complement, antibodies against complement and other triggering factors for endothelial dysfunction, until a new therapeutic agent was found.

In a study by Fuge et al.,¹⁹ six out of 17 post-BMT TMA patients responded to PE, but just one of these six patients survived, resulting in a 6% success rate. Roy et al.⁹¹ described 17 post-BMT TMA patients treated with PE during a 10 year period, 1989–1998. Three months after diagnosis, four out of 17 patients were alive. However, just one patient had survived by 2001, the year in which they published the report, resulting in a success rate again of 6%. PE may confer some benefit in patients with secretory antibodies. However, PE can theoretically potentiate the effects of TMA by complement activation.^92,93

In 2011, eculizumab, the humanized monoclonal antibody, became the only therapy approved by the Food and Drug Administration for the management of aHUS. Eculizumab inhibits activation of the terminal complement pathway (C5), and its effectiveness in treating aHUS was demonstrated in many studies.^94–97

Legendre et al.⁹⁶ studied the outcome of 37 aHUS patients after 26 weeks of eculizumab therapy in two clinical trials. In total, 80% of patients treated with eculizumab had improvements in anemia, thrombocytopenia, and renal function, resulting in eventual cessation of dialysis.

In a recent study, Baskin et al.⁹⁸ studied retrospectively 15 children diagnosed with aHUS. Following the initiation of eculizumab, all patients achieved full recovery of renal function and hematological parameters.

Jodele et al.⁹⁵ treated six children with severe post-BMT TMA using eculizumab. They observed that post-BMT TMA resolved over time in four out of six children after achieving therapeutic eculizumab levels. They also mentioned that to achieve therapeutic drug levels and a clinical response, children with post-BMT TMA required higher doses or more frequent eculizumab infusions than currently recommended for children with TMA.

Early initiation of eculizumab therapy may have great therapeutic success with respect to improving renal function and reversing damage in post-BMT TMA patients.⁹⁴ Future studies will be required to assess its efficacy more thoroughly in this patient population.