The proband (Subject A-V-23; Fig. 1A) presented at age twenty-five with a ten-year history of progressive pancytopenia (Fig. 2A). We have previously found that he was heterozygous for a loss-of-function TERT K570N mutation [15]. Eighteen members of his immediate family (Fig. 1A, inset) were previously screened and eight tested positive for the mutation [15]. A long history of hematologic diseases was well known in the family back to the patient's paternal great-great-grandmother, Subject A-I-2, who died of an apparent blood disorder at the age of sixty-five. However, the great-grandmother and the grandfather, Subjects A-II-7 and A-III-16 (the latter previously found to be positive for the mutation [15]), had not manifested hematologic symptoms. The proband's father (Subject A-IV-29) had thrombocytopenia and mild anemia from childhood. By age thirty-three, he developed myelodysplasia, which rapidly progressed to acute myeloid leukemia and death subsequent to an induction cycle of chemotherapy (Fig. 2B). The father was an obligatory carrier, as three of his sisters and his father also tested positive and his wife (the proband's mother) tested normal. Another of the proband's heterozygous paternal aunts (Subject A-IV-26) and the index patient's two heterozygous sisters, Subjects A-V-19 and A-V-20, ages forty-seven, twenty-two, and nineteen, respectively, only have macrocytosis, whereas his two wild-type sisters (Subjects A-V-21 and A-V-22) are healthy and without hematologic abnormalities. The patient's eldest of three sons, now six years old, carries the mutation, but he is asymptomatic and has normal blood counts.

The family has a history of severe liver disease. In our first report [15], only one paternal aunt (Subject A-IV-23) was found to carry the mutation and have liver disease. She underwent successful liver transplantation at age twenty for a non-A, non-B hepatitis that rapidly evolved to submassive hepatic necrosis with early fibrosis (Fig. 2C). Pathological examination of her liver uncovered massive necrosis without significant hepatitis, which is not specific for a particular etiology. Masson stain revealed early fibrosis in areas of parenchymal collapse and at the edges of portal areas and around central veins. She experienced anemia during pregnancy, but she otherwise had no history of hematologic abnormalities. The paternal aunt with a twenty-year history of aplastic anemia (Subject A-IV-25) and also heterozygous for the mutation developed dyspnea and cough at the age of forty-six after our first report. Spirometry revealed a moderately restrictive pattern and a very severe diffusion defect. Chest computed tomography showed heterogeneous bilateral peripheral lower lung infiltrates, suggestive of pulmonary fibrosis. Prednisone therapy was initiated, but she developed rapidly accumulating ascites, and was diagnosed with non-cirrhotic portal hypertension. She tested negative for hepatitis-associated viruses. Hepatocellular and canalicular enzymes were in the normal range, but her albumin was mildly low (3.6 g/dL) and total bilirubin was mildly increased (1.6 g/dL). Liver biopsy revealed no fibrosis connecting portal areas or perisinusoidal fibrosis. However, portal areas were small without visible vein (Fig. 2D). Hepatocytes showed variation in cell and nuclear size, and varying plate width, consistent with regeneration on reticulin stain (Fig. 2E). CD34 abnormally stained in the sinusoidal endothelial cells around the portal areas and central veins, consistent with an abnormal proportion of arterial blood flow to the sinuses (Fig. 2F). Iron was heavily accumulated, mainly in hepatocytes in zone 1. Her pulmonary function rapidly deteriorated, and she died of respiratory insufficiency.

We now expanded the genetic screening to an additional 35 relatives (Fig. 1A). Only one relative tested positive for the mutation, which was a first-cousin-twice-removed (Subject A-III-11) who died at the age of forty-eight with the diagnosis of liver cirrhosis. Her mutational status was demonstrated by sequencing DNA extracted from a paraffin-embedded liver biopsy obtained from hospital archives. She presented with pyoderma gangrenosum and upon workup, her liver enzymes and bilirubin were elevated (AST, 124 IU/L; alkaline phosphatase, 410 IU/L; albumin, 3.8 g/dL; total bilirubin, 3.8 mg/dL), and liver biopsy (Fig. 2G) demonstrated pre-cirrhotic chronic cholestatic liver disease with a differential diagnosis of primary sclerosing cholangitis or a sclerosing cholangitis secondary to autoimmune hepatitis. Serology for hepatitis A and B were negative at the time. She was treated with azathioprine and prednisone for one month, after which time she became severely jaundiced. Her clinical state deteriorated, and she died of fungemia. Her mother, a great-great-aunt to the index patient (Subject A-II-4) died at forty-nine of liver cirrhosis. Unfortunately, no biopsy specimen or clinical records were available. Because her husband was not related to the proband, and her daughter (A-III-11) carried the mutation, she was an obligatory carrier (Fig. 1A).

There was no nail dystrophy, leukoplakia, or skin hyperpigmentation, all physical features characteristic of dyskeratosis congenita, in any of the mutation-carriers. Although the index patient and some of his relatives showed a strikingly premature graying of hair, surprisingly this characteristic did not track with the mutation. No history of alcohol consumption or smoking was present in the family.

The proband (Subject B-II-3; Fig 1B), a 57-year-old man heterozygous for a novel TERC nucleotide 341–360 deletion not found in 188 controls, presented a six-year history of Barrett's esophagus and recently worsening dysphagia. Imaging displayed a tumor at the gastro-esophageal junction, which was identified as esophageal cancer by biopsy. During the initial evaluation, the patient was pancytopenic, and his bone marrow revealed 5% cellularity with trilineage hypoplasia but without malignant infiltration. His liver enzymes and function tests also were abnormal (alkaline phosphatase, 482 IU/L; albumin, 2.2 g/dL; total bilirubin, 0.9 mg/dL). The tumor was surgically resected, and concurrent liver biopsy showed cirrhosis with foci of lobular inflammation dominated by plasma cells, extensive sinusoidal fibrosis, and Mallory bodies (Fig. 2H). Serological tests for hepatitis-associated viruses were negative. His father (Subject B-I-1) had suffered from gastroesophageal reflux disease and died at age eighty of a poorly differentiated adenocarcinoma at the gastroesophageal junction with hepatic metastasis. The proband's brother (Subject B-II-2) also had gastroesophageal reflux disease and Barrett's esophagus. No pathological specimens were available from his father and his brother for genetic screening. His brother's son (Subject B-III-4) was reported to abuse alcohol and to have cirrhosis, but he tested negative for the mutation. The proband's thirty-two year-old son (Subject B-III-7) is heterozygous for the TERC deletion; he had mild macrocytic anemia during a routine medical visit. Upon further investigation, he was found to have an enlarged fatty liver on ultrasound though his liver enzymes were normal; his bone marrow biopsy was hypocellular (20%), and his liver biopsy showed macrovesicular steatosis with foci of lobular inflammation, portal chronic inflammatory infiltrate, and mild hepatocellular iron accumulation (Fig. 2I). Tests for hepatitis B and hepatitis C viruses were all negative. Pulmonary function test results were within normal limits. He has an eleven-year history of social alcohol consumption.

The proband (Subject C-III-1; Fig. 1C), a thirty year-old Caucasian male previously fond to be heterozygous for a TERC nucleotide 28-34 deletion [15], had a thirteen-year history of moderate aplastic anemia. His father (Subject C-II-2) had a long history of thrombocytopenia and died at age thirty-two of fungal sepsis; autopsy revealed mixed micro and macronodular liver cirrhosis, chronic congestive splenomegaly, esophageal varices, diffuse interstitial pulmonary fibrosis with mild chronic inflammation, and a mildly hypoplastic bone marrow (50%). The index patient's paternal uncle (Subject C-II-1) had died of myelodysplasia. We now screened his brother (Subject C-III-3), with a long history of mild pancytopenia and elevated liver enzymes, and he also tested positive for the TERC deletion. Liver biopsy demonstrated hepatocytes with mild variation in nuclear size, mild hepatocellular iron accumulation in a pericanalicular pattern (Fig. 2J), and several zones displaying abnormally widened hepatocyte plates, consistent with regeneration (Fig. 2K).

The proband, a thirty-eight year-old female, was found to be heterozygous for a novel TERC nucleotide 109–123 deletion not observed in 188 controls. She had a six-year history of transfusion-independent pancytopenia first detected during pregnancy. At that time, her hemoglobin was 6 g/dL, and her anemia was unresponsive to erythropoietin treatment. Her bone marrow biopsy was 5% cellular with trilineage hypoplasia and a transient clonal chromosome 1 abnormality [der(1) t(1;1)(p36.2;q;12)]. Her wild-type mother (Subject D-II-2) had history of resolved anemia in the past but is otherwise healthy. Her father (Subject D-II-1), a probable carrier as her mother tested negative, had a 15-year history of hepatitis and liver cirrhosis and died at the age of forty-five years of massive gastrointestinal bleeding. At necropsy, the liver was cirrhotic and microscopically showed moderate fatty change and hyaline Mallory bodies; spider angiomata, jaundice, ascites, and esophageal varices also were present. He had a history of moderate alcohol consumption. Unfortunately, no pathologic specimen was available for further analysis or genetic testing. The index patient's paternal grandfather, (Subject D-I-1) also died of cirrhosis at a young age (pathologic specimens were not available).

The proband (Subject E-III-3), a fifteen-year-old male, presented in 1992 with a history of hemorrhage. Laboratory tests revealed pancytopenia and elevated alkaline phosphatase; liver function tests were within normal limits. Bone marrow biopsy showed 25% cellularity with normal cytogenetics. He was treated with androgens without much benefit for his blood counts. As his hematological status deteriorated, he underwent an unrelated hematopoietic stem cell transplant but died of a transplant-related complication. His father (Subject E-II-1) was forty-four years old when first seen, and at the time, he had a 10-year history of thrombocytopenia and leukopenia along with a mildly hypocellular bone marrow. This unusual association between aplastic anemia and liver cirrhotic disease observed in this pedigree and in an additional family led us to describe a “new familial syndrome” in 1997, which appeared to have an autosomal dominant inheritance, but genetic analysis was not available at that time (family E in the present series corresponds to family A in our previous report) [16]. Six years post-presentation, the father developed a nonproductive cough and dyspnea on exertion. Spirometry revealed a reduced diffusion capacity and computed tomography of the chest was consistent with pulmonary fibrosis. During evaluation, some liver enzymes were elevated, and liver biopsy findings were consistent with hepatoportal sclerosis complicated by nodular regenerative hyperplasia. He had no history of ethanol consumption or smoking. Microscopic examination revealed portal areas with chronic inflammatory infiltrate but with no interface hepatitis (Fig. 2L). The portal veins were either missing or slit-like in most of the portal areas. Hepatic architecture was subtly distorted by nodularity with zones of small compressed hepatocytes alternating with zones of large hepatocytes with widened plates (Fig. 2M). CD34 staining was abnormally positive in sinusoidal endothelial cells, mainly around the portal areas and central veins (Fig. 2N). There was sinusoidal dilatation and congestion near central veins. Iron was accumulated within hepatocytes in zones one and two. Ultimately, the respiratory symptoms evolved, and the patient died of respiratory insufficiency. Serological tests for hepatitis B and hepatitis C viruses were negative. Sixteen years post-presentation, the DNA extracted from his paraffin-embedded liver specimen revealed a novel heterozygous TERT S368F mutation not present in 528 healthy controls. The proband's paternal grandfather (Subject E-I-1) died at age sixty-four, and autopsy revealed pulmonary fibrosis, nodular hyperplasia of the liver, and splenomegaly. A paternal aunt (Subject E-II-3) died at age thirty-seven with extensive pulmonary fibrosis and liver cirrhosis. She presented macrocytosis in peripheral blood but normocellular bone marrow. A paternal uncle (Subject E-II-4) died at age twenty-three with massive gastrointestinal bleeding and thrombocytopenia. Autopsy revealed splenomegaly with expanded red pulp, consistent with portal hypertension, and the liver appearance was consistent with portal fibrosis but not cirrhosis. The other paternal uncle (Subject E-II-5) died at age thirty-five years, and autopsy revealed aplastic anemia, splenomegaly, macronodular cirrhosis, and portal fibrosis. No samples were available for the genetic testing of the other affected individuals, including the proband. The proband's two siblings (Subjects E-III-1 and E-III-2), mother (Subject E-II-2), and cousin (Subject E-III-4) appear to be healthy, and each tested negative for the mutation.

Further details of the clinical cases described here are available to other researchers on request.

In the three generations of Family A analyzed by flow-FISH, mutation carriers (Subjects A-III-16, A-IV-23, A-IV-25, A-IV-26, A-V-19, A-V-20, and A-V-23) had total peripheral blood white cell telomere lengths below the shortest percentile of healthy age-matched controls (Fig. 3A). In contrast, the wild-type individuals analyzed (Subjects A-IV-28, A-V-21, and A-V-22) were found to have telomere lengths between those of their heterozygous family members and the 50th percentile. In the additional four families studied, all tested mutation carriers had lymphocyte and neutrophil telomere length below the shortest percentile (Fig. 3A). Interestingly, four non-carriers in families A and E, each of whom had a heterozygous parent, also had short telomeres for their age (Fig. 3A).

We previously showed that TERT K570N results in a complete loss of the ability of telomerase to add hexameric repeats to telomeres and that TERC nucleotide 28–34 deletion reduces telomerase enzymatic activity - each by haploinsufficiency [15]. Here we found by co-transfection experiments that TERC nucleotide 109–123 deletion completely abolished telomerase enzymatic activity, TERC nucleotide 341–360 deletion reduced telomerase activity to approximately one-third of that observed for wild-type TERC, and TERT S368P reduced telomerase activity to approximately 10% of that observed for wild-type TERT (Fig. 3B).

Patients and their relatives or guardians of minors provided written informed consent for genetic testing, according to protocols approved by the institutional review board of the National Heart, Lung, and Blood Institute, protocol 04-H-0012 (www.ClinicalTrials.gov identifier: NCT00071045). Clinical investigation was conducted according to the principles expressed in the Declaration of Helsinki. Continuing review application of the protocol was last renewed by NHLBI IRB in February 26^th, 2009; the protocol was last amended on September 9^th, 2009. Patients who are described in detail have read the manuscript and provided written informed consent for publication.

The probands of each family were referred to the National Institutes of Health Hematology Branch clinic for evaluation of bone marrow failure, and they were diagnosed with aplastic anemia based on conventional bone marrow and blood-count criteria [33]. In Family A, relatives were invited for clinical and genetic evaluation and responded a questionnaire regarding their health status and specifically addressing hematologic, immune, pulmonary, and hepatic diseases. When a clinical history was positive, subjects were requested to provide clinical tests and medical records for further analysis. Medical records also were obtained from deceased relatives who had had liver or pulmonary disease, upon family approval. In the other families, relatives were invited for clinical evaluation at the NIH Clinical Center and medical records from outside institutions were obtained after consent; medical records also were obtained from deceased relatives, upon family consent. Complete blood counts were performed at the NIH Clinical Center for clinically healthy subjects found to carry a telomerase mutation, and the diagnosis of macrocytosis or anemia were established based on the standard laboratory parameters. Patients and their relatives or guardians of minors provided written informed consent for genetic testing, according to protocols approved by the institutional review board of the National Heart, Lung, and Blood Institute. DNA was extracted from peripheral-blood or buccal mucosal cells as previously described [34]; for Subject A-III-11 and Subject E-II-1, DNA was obtained from paraffin-embedded liver tissue (PicoPure DNA Extraction Kit, Arcturus, Mt View, CA).

DNA samples from 188 healthy persons served as controls for TERC and TERT gene mutations: 117 were white (94 from Human Variation Panel HD100CAU, Coriell Cell Repositories [http://locus.umdnj.edu/nigms/cells/humdiv.html], and 23 from SNP500Cancer [http://snp500cancer.nci.nih.gov]), 24 black (from SNP500Cancer), 23 Hispanic (from SNP500Cancer), and 24 Asian (from SNP500Cancer) [35]. An additional 340 healthy controls were screened for TERT gene mutations: 94 blacks from Human Variation Panel HD100AA and 246 anonymous healthy subjects of Hispanic origin (52 percent Peruvians, 28 percent Latin Americans, and 20 percent Pima and Maya Amerindians) [10].

TERT and TERC genotyping was performed as previously described [10].

Telomere length of peripheral blood leukocytes was measured after red cell lysis with ammonium chloride solution by flow cytometry-fluorescent in situ hybridization (flow-FISH) as reported previously [15], [36].

Wild-type vector was mutagenized by Mutagenex, and sequence was confirmed by direct sequencing of the whole insert, and plasmids were purified using the HiSpeed Plasmid Maxi Kit (Qiagen). Vectors containing disease-associated telomerase mutations were transfected into VA13 cells, and telomerase activity measured as previously described with slight modifications [10]. In the present study, we used a fluorescent telomere repeat amplification protocol (TRAP), XL TRAPeze assay (Chemicon) and fluorescence was measured in a Victor 3 multilabel plate reader (Perkin Elmer). Telomerase activity was measured in total product generated (TPG) units based on the standard curve results and calculated strictly according to the manufacturer's manual and expressed as telomerase activity relative to wild-type TERT and TERC, which was considered 100%.