Toxic leukoencephalopathy caused by chemotherapeutic drugs other than methotrexate

Background & Objective Chemotherapy-induced toxic leukoencephalopathy is clinically characterized by progressive cognitive loss, often resulting in sudden death. The objective of this study is to share distinctive clinicopathological features of chemotherapy-induced brain change. Methods The brains of a 64-year-old woman and a 63-year-old man who suffered from rapid deterioration of consciousness were autopsied. The initial clinical impressions were central nervous system (CNS) graft versus host disease (GVHD), infectious or autoimmune encephalitis. Both patients had been treated with multidrug chemotherapy, including cytarabine arabinoside, daunorubicin, udarabine, azacitidine, and busulfan, and allogeneic peripheral blood stem cell transplantation because of hematological malignancy (acute myelogenous leukemia and myelodysplastic syndrome). Results The autopsies revealed a vacuolar change of the white matter with axonal spheroids, reactive gliosis, and foamy macrophage inltration in the brain, predominantly in the visual pathway of the occipital and temporal lobes. There was no lymphocytic inltration in the brain tissue, which is characteristic of CNS-GVHD or encephalitis, suggesting these syndromes were not the cause of brain illness. cases methotrexate our autopsy cases also drugs other than methotrexate. reported two cases of autopsy-proven leukoencephalopathy caused by chemotherapy other than methotrexat. The limitation of our study is that we do not know exactly which drugs cause signicant change in the brain, and have not found the underlying mechanism of chemobrain. In this way, we were able to share distinctive clinicopathologic ndings and caution other from misdiagnosing and mistreating patients with similar cases. Autopsies revealed a vacuolar change of the white matter with many axonal spheroids and reactive gliosis. Scattered foamy macrophage inltration was found predominantly in the visual pathway of the occipital and temporal lobes. Staining for CD68, TMEM119, CD3, CD8, and CD20 revealed no lymphocytic inltration in the entire brain tissue. However, exogenous foamy macrophage inltration was present. The clinicopathologic features of these two cases identical chemotherapy-induced leukoencephalopathy cases. Autopsies are a medical to determine uncertain of However, evidence-based of could be a an our of chemo


Page 3/16
confusion, somnolence, ataxia, spasticity, and major seizures. These progressed to dementia, coma, and death. Rubinstein et al. reported characteristic pathological ndings of DNL represented by multiple irregular foci of demyelination with spongiosis. In a severe case, acellular coagulative brain necrosis presented with reactive astrocytosis. The pathology was prominent in the white matter of the cerebrum, cerebellum, and brainstem. Furthermore, axonal spheroids were evident in the lesions [2].
In 2009, Matsubayashi et al. introduced a new type of chemotherapy-induced disseminated demyelinating leukoencephalopathy (DDL). Macroscopically, DDL is characterized by extensive fragility of the brain. Microscopically, it has widespread, demyelinated foci without signi cant axonal changes.
These were distributed especially in the occipital lobe compared with other lobes [3].
In 2011, The European Association of Neuro-Oncology emphasized that neurological complications of chemotherapy need to be separated from central nervous system (CNS) dysfunctions caused by tumors themselves. This includes dysfunction caused by tumor metastasis, infection, or metabolic changes.
Complete or partial recovery may occur after chemotherapy is completed; however, progression to irreversible damage and even death were possible outcomes as well. It is crucial to understand the toxic effects of chemotherapy on the brain and, subsequently, identify means to prevent them. The European association de ned chemotherapy-induced neurotoxic complications and listed the respective causative agents. They also classi ed clinical syndromes caused by cytostatic drugs into 11 categories: acute (reversible) encephalopathy, subacute encephalopathy, chronic encephalopathy, posterior reversible (leuko-) encephalopathy syndrome (PRES), multifocal leukoencephalopathy, thrombotic microangiopathy, cerebral infarctions, cortical blindness, cerebellar dysfunction, seizure, and aseptic meningitis. PRES is clinically characterized by headaches, visual disturbances, confusion, seizures, and eventually coma.
We report two cases of autopsy-proven PRES that were irreversible and caused by various chemotherapy regimens other than methotrexate, including cytarabine, daunorubicin, udarabine, and busulfan, and cytarabine arabinoside [Ara-C]. This study highlights the distinctive clinicopathological features which may help clinicians from misdiagnosing and mistreating patients with similar cases.

Material And Methods
The brain-only autopsies were conducted in two deceased patients with routine methods. The clinical differential diagnosis of the patients' brain illnesses included infectious or autoimmune encephalitis or central nervous system-graft versus host disease (CNS-GVHD), or toxic or metabolic encephalopathy.
Hematoxylin and eosin (H&E) staining and immunohistochemical analysis were carried out on formalinxed, para n-embedded blocks. Primary antibodies for immunohistochemistry included NeuN

Result
Case presentation and imaging study Autopsy performed patients were two deceased adults (64y/female and 65y/male), who shared the history of treating various chemoregimen and peripheral blood transplantation. A 62-year-old woman with acute myeloid leukemia had been administered chemotherapy, including Adriamycin, cytarabine arabinoside (Ara-C), daunorubicin (AD 7 + 3) and Adriamycin, Ara-C, idarubicin (AI 7 + 3). However, leukemia recurred 20 months after cessation of treatment, and she was diagnosed with acute undifferentiated leukemia. She was treated with FLAG induction chemotherapy ( udarabine, Ara-C, granulocyte colony-stimulating factor) and had achieved a favorable response. Intermediate-dose Ara-C #1 (IDAC) was administered for consolidation. She received a full match allogeneic hematopoietic stem cell transplant from a male donor, being registered in Korea Marrow Donor Program (KMDP). She had taken tacrolimus as GVHD prophylaxis. One month after transplantation, she became drowsy.
Mood disorder and hypoactive delirium appeared initially. The patient progressed to a subacute and chronic altered mental state. Cerebrospinal uid (CSF) examination revealed polymorphonuclear neutrophil dominant pleocytosis and increased protein level up to 135 g/dL (cf. normal range: 15-45 g/dL). Three months after transplantation, the patient's consciousness deteriorated rapidly, but light re exes were still intact. However, she did not respond to painful stimulations and was unable to communicate.
T2-weighted magnetic resonance imaging (MRI) showed high signal intensity in the bilateral centrum semiovale, and optic pathway with slightly decreased size of bilateral basal ganglia (Fig. 1). GVHD, toxic or metabolic encephalopathy, or encephalitis was considered as the differential diagnosis. The patient was treated with methylprednisolone and intrathecal steroid for suggested CNS-GVHD. In addition to neurological symptoms, she developed respiratory failure and shock, which appeared to be caused by sepsis and aspiration pneumonia despite aggressive use of broad-spectrum antibiotics. X/Y uorescence in situ hybridization of the peripheral blood revealed 0.2% of the XX chromosome of the host and 99.8% of the XY chromosome of the male blood donor. The short tandem repeat examination showed her remnant host blood cells was 5.6%. She rapidly deteriorated, and an autopsy was performed to determine brain pathology.
The second patient (64y/male) had been diagnosed with myelodysplastic syndrome (MDS) with alcoholic liver cirrhosis and histories of multiple transfusions for anemia at other hospitals. He had been treated with bridging azacytidine #3 for MDS for three months and received busulfan-udarabine-antithymocyte globulin (BuFluATG) conditioning. He received a full match hematopoietic stem cell transplant from a male donor, being registered in the KMDP. He received tacrolimus (for 13 days) as CNS-GVHD prophylaxis. One month after transplantation, he displayed general weakness, gait disability, and fecal and urinary incontinence. Forty days after transplantation, he became drowsy and rapidly lost consciousness. After 10 days, he was unable to maintain eye contact and had a cognitive disorder with a sleep tendency. Brain MRI revealed a restrictive lesion on the left splenium of the corpus callosum, multiple T2 high signal intensity lesions at the central pons, and bilateral parietal periventricular white matter ( Fig. 1).
Based on MRI ndings and increased CSF protein levels (81 g/dL), CNS-GVHD or CNS infection was suggested. Steroid (methylprednisolone 1g/day for 5 days), intravenous immunoglobulin (IVIg) (400 mg/kg for 5 days), and Keppra [500 mg bid (twice a day) IV] were administered, and he recovered the ability to speak his name. However, he soon fell into coma. CNS-GVHD was suggested, and physicians started tacrolimus IV 0.04 mg/kg/day, targeting a dose of 10-20 ng/mL. Thereafter, methylprednisolone, tacrolimus, mycophenolate mofetil, intrathecal steroid, rituximab, and ruxolitinib were administered to suppress immune reaction. His consciousness worsened, and he presented with coarse respiratory sounds. Blood culture tested positive for extended-spectrum beta-lactamases (ESBL) + Klebsiella. He was diagnosed with ESBL-positive Klebsiella pneumonia and septic shock with disseminated intravascular coagulopathy. Although vancomycin was added to the broad-spectrum antibiotics regimen, he succumbed to multiorgan failure four months after transplantation.

Autopsy and pathological ndings
The primary clinical information and neuropathological ndings are summarized in Table 1 and Figures.
Brain autopsy of case 1 revealed symmetric global atrophy, pale parenchyma, and pale substantia nigra ( Fig. 2). The brain weighed 1060 grams, and CSF was clear. Atherosclerosis was found in a short segment of the basilar artery.
Spongiform (vacuolar) change of the white matter highly suggested chemotherapy-related brain damage. Axonal spheroids, foamy macrophage in ltration, and reactive gliosis in the white matter, especially in the optic pathway of the occipital and temporal lobes, all consistent with previously reported chemotherapyinduced brain change. The nal diagnosis was chemotherapy-induced PRES (Fig. 3). CD68-positive and TMEM-negative foamy macrophages were in ltrated in the cerebral white matter and the optic pathway of the occipital and temporal lobes. However, TMEM119-positive microglia were markedly reduced in the gray matter and mildly reduced in the white matter of the occipital lobe.
Mild white matter rarefaction was identi ed in the centrum semiovale with LFB and MBP stains. Many βamyloid-positive diffuse-type senile plaques, but no neuritic plaques, were found in the neocortex, putamen, subiculum, and entorhinal cortex [Thal phase 2, consortium to establishe a registry for Alzheimer's disease (CERAD) stage 0). H&E stain and synaptophysin immunohistochemistry of the globus pallidus showed severe neuronal loss. Focal loss of dopaminergic neurons was identi ed in the midbrain, but AT8, α-synuclein and p-TDP43 were totally negative in the brain tissue, suggesting no other neurodegenerative disorders except a primary age-related tauopathy, (PART) which was con rmed with AT8 immunohistochemistry in the entorhinal cortex. There was no evidence of cerebral amyloid angiopathy. Iron stain revealed no iron deposit in the entire brain, including the basal ganglia. Gomori's methanamine silver and Periodic acid Schiff stains revealed no fungal organisms. There was no evidence of viral or autoimmune encephalopathy.
We carefully investigated possible donor-derived in ammatory cell in ltration. However, there was no in ammatory cell in ltration on H&E stain or CD3, CD8, and CD20 immunostaining in the entire brain, which suggested against CNS-GVHD.
Brain autopsy of case 2 showed no gross abnormaltity, and the brain weighed 1310 grams. Atherosclerosis and arteriosclerosis were not identi ed. However, H&E, GFAP, NF, LFB, CD68, and TMEM119 stains revealed vacuolization of the white matter with diffuse axonal spheroids, reactive gliosis, and foamy macrophage in ltration in the cerebral white matter and the optic pathway of the temporal and occipital lobes. This was identical to that of case 1. GFAP revealed reactive gliosis of the entire neocortex. In CA1 and CA2 of the hippocampus and entorhinal cortex, many pTau positive neuro brillary tangles and tau neuropil threads (Braak stage 2) were found, which was consistent with PART, but there was no evidence of other neurodegenerative disorders when analyzed immunohistochemically with AT8, α-Synuclein, β-amyloid, and TDP43.
TMEM119 was positive in the host microglia. Similar to case 1, there was no lymphocytic in ltration in the entire brain seen with CD3, CD8, and CD20 immunostaining, which suggested against a GVHD diagnosis. The nal autopsy diagnosis was Chemotherapy-induced toxic leukoencephalopathy with predominant optic pathway involvement.

Discussion
Systemic chemotherapy, often with multidrug combinations, is a treatment of choice for oncological malignancy, especially for patients with hematopoietic cell malignancy. "Chemotherapy-induced toxic leukoencephalopathy" is currently a rising clinical syndrome characterized by progressive cognitive disorder and detrimental effects on the quality of life, which may lead to sudden death. However, a few data have been accumulated and most are methotrexate-induced leukoencephalopathy.
Here, two cases of autopsy-proven toxic leukoencephalopathy caused by chemotherapy regimens other than methotrexate were reported. In this way, we are able to share distinctive clinicopathologic ndings and caution other physicians to make accurate diagnosis and treatment when meeting similar patients. Both patients were administered multiple courses of various chemotherapy regimens, in which udarabine was included. Clinically, both patients presented with a neurological disorder, including cognitive disorder and motor and sensory loss, which rapidly progressed to coma and death. Brain autopsies depicted cerebral white matter spongiform change, axonal spheroids, and foamy macrophage in ltration. In ltrating macrophages were CD68-positive but TMEM119-negative in both cases. CD3, CD8, and CD20 stains revealed no lymphocytic in ltration in both cases. No apoptotic cells were identi ed. Pathologic ndings of these two cases were consistent with Chemotherapy-induced toxic leukoencephalopathy. [4][5][6][7][8][9][10] Methotrexate is the most common drug associated with chemo brain, characterized by acute (reversible) encephalopathy, subacute encephalopathy, chronic encephalopathy, cerebral infarctions, seizure, or aseptic meningitis. [3] However, other chemotherapy drugs are also associated with toxic encephalopathy symptoms that present with the above-mentioned symptoms. Chronic encephalopathy and PRES are related to high dose multi-chemotherapies, including cyclophosphamide, Ara-C, cis-platinum, ifosfamide, vincristine, gemcitabine, and other immunosuppressants. [9,11] Chronic encephalopathy usually develops after a latency of some months to years, often presenting progressive and irreversible clinical manifestations. PRES is clinically characterized by headaches, visual disturbances, confusion, seizures, and eventually coma.
In 2004, Lai et al. reported an autopsy-proven, methotrexate-based, chemotherapy-induced leukoencephalopathy in primary CNS lymphoma [12]. Unfortunately, treatment-related leukoencephalopathy is the leading brain pathology after the successful treatment of primary CNS lymphoma (PCNSL). Lai et al. reviewed ve more autopsied patients who died of leukoencephalopathy [12]. Neurological symptoms developed at a median of 1 month after treatment completion. The median survival was 30 months (range, 22-68 months) after neurotoxicity onset. All had hyperintensity on T2weighted MRI, and two patients presented with enhancing lesions that were observed 5 and 14 months after the treatment, respectively. The autopsy revealed no residual PCNSL. Common pathologies were myelin and axonal loss, reactive gliosis, spongiosis, and rarefaction of the white matter. Two patients had brain necrosis, which correlated to the enhancing lesions seen on MRI. Interestingly, all had small vessel disease, and four had atherosclerosis of large cerebral vessels in the circle of Willis; two had recent strokes that were discovered during autopsy. The authors concluded that Chemotherapy-induced toxic leukoencephalopathy is not always a late or delayed consequence of chemoradiation therapy but also can develop very early in some patients. They also suggested that vascular disease may be a component of this injury [12].
Our cases might be due to udarabine, Ara-C, busulfan, and cytarabine treatment. High dose udarabine may cause cortical blindness, which has been associated with the administration of immunosuppressive drugs, antibodies, and other substances [4]. Ara-C can induce cerebellar dysfunction and aseptic meningitis [9,12]. Busulfan, cyclosporine, vincristine, cis-platinum, methotrexate, and paclitaxel may cause seizures [4]. The frequency and severity of CNS toxicity depend on the drug, cumulative doses, the duration of treatment, and additional risk factors such as coexisting neurological morbidity [13]. Wellknown factors that increase the risk are dose escalation, combination therapy, stem cell transplantation, and irradiation of the brain [9].
In 1994, Cheson et al. reviewed the neurotoxicity of purine analogs that are widely used in indolent lymphoid malignancies, including udarabine, cladribine, and pentostatin [14]. They compared the adverse drug effects of udarabine in chronic lymphocytic leukemia and cladribine and pentostatin in hairy cell leukemia. The neurotoxicity spectrum of these drugs was similar, which include myelosuppression, immunosuppression, and sporadic neurotoxicity [14]. All three drugs lead to lifethreatening or fatal neurotoxicity at higher-than-recommended doses. Each agent-induced neurologic complication occurs in approximately 15% of patients at the recommended doses, mostly mild and reversible. However, severe neurologic de cits were encountered. They were occasionally delayed, often at least partially reversible, or sometimes fatal [14].
In 1994, Zabernigg rst reported about late-onset fatal neurotoxicity induced by low dose udarabine monotherapy in patients with B-cell chronic lymphocytic leukemia (CLL) [10]. A 55-year-old man developed a severe neurological disorder six months after nishing six cycles of udarabine monotherapy. He presented with aphasia, apraxia, acalculia, hemihypesthesia, and spastic hemiparesis. A computed tomography scan of the brain showed multiple low-density areas involving the subcortical white matter. MRI showed subcortical white-matter abnormalities compatible with demyelination. Bizarre astrocytes and swollen oligodendrocytes were noted in the subcortical area. Multinuclear inclusion bodies were identi ed in some oligodendrocyte nuclei. The pathologic diagnosis was progressive multifocal leukoencephalopathy. Finally, he succumbed to coma and death [10]. Although a low dose of udarabine was used, the spectrum and severity of neurotoxicity were not different from that of a high dose.
The underlying cellular mechanism is still unclear. However, some models based on clinical and animal experiments help us speculate the possible mechanism of chemo brain. Peripheral cytokines initiate the development of chemo brain [15][16][17]. This cytokine-mediated signaling cascade induces persistent epigenetic alterations. These epigenetic changes alter gene expression, metabolic activity, and neuronal transmission that ultimately affect cognitive function. Chemotherapy drugs cause cellular stress and injury [15][16][17]. This induces an in ammatory response in the periphery, releasing cytokines such as tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), IL-8, IL-10, and monocyte chemoattractant protein-1 (MCP-1). Peripheral cytokines can access the brain through leaky lesions in the blood-brain-barrier or can be transported via active mechanisms. [18,19] It is assumed that peripheral cytokines communicate with the cytokines of the CNS through the local in ammatory network. [20] Peripheral cytokines stimulate endothelial cells and perivascular macrophages, monocytes, and T cells in the brain to produce similar local cytokines and chemokines.
[18] Microglia, astrocytes, oligodendrocytes, and neurons respond by releasing additional cytokines and chemokines in the brain. As a consequence of this cascade, oxidative stress increases, neurogenesis and neuroplasticity decreases, and neuronal excitotoxicity increases.
[16] These factors ultimately in uence neurotransmitters and neuronal alteration [16, 21,22]. Cytokines in the CNS are known to be mainly derived from microglia. [16,19] Along with the strong evidence of correlating peripheral cytokines and cognitive dysfunction, microglial activation has been identi ed as a key factor in the reactivation of astrocytes and dysfunctional oligodendrocyte precursor cells in the previous work.
[23] TMEM119 is a reliable microglial marker, as it distinguishes microglia from circulating macrophages that ow into the brain [24]. Recently, neuroimaging studies have broadened our understanding of the structural and functional changes in progressive cognitive dysfunction, showing a reduction in the frontoparietal white and gray matter [13,25].

Conclusion
We reported two cases of autopsy-proven leukoencephalopathy caused by chemotherapy other than methotrexat. The limitation of our study is that we do not know exactly which drugs cause signi cant change in the brain, and have not found the underlying mechanism of chemobrain. In this way, we were able to share distinctive clinicopathologic ndings and caution other physicians from misdiagnosing and mistreating patients with similar cases. Autopsies revealed a vacuolar change of the white matter with many axonal spheroids and reactive gliosis. Scattered foamy macrophage in ltration was found predominantly in the visual pathway of the occipital and temporal lobes. Staining for CD68, TMEM119, CD3, CD8, and CD20 revealed no lymphocytic in ltration in the entire brain tissue. However, exogenous foamy macrophage in ltration was present. The clinicopathologic features of these two cases were identical to previously published chemotherapy-induced leukoencephalopathy cases. Autopsies are a medical research method to determine the uncertain cause of death. However, if no evidence-based report is garnered, the cause of death could be a mystery, even though an autopsy has been conducted. Hence, our report of chemo brain is valuable.