Introduction

Hepatic venous outflow tract obstruction (HVOTO) and extrahepatic portal venous obstruction (EHPVO) are important causes of portal hypertension in the Indian subcontinent [1, 2]. These conditions are related to blockage of the hepatic veins and/or inferior vena cava (IVC) in HVOTO and of the portal vein in EHPVO. This blockage is most often due to venous thrombosis, though occasionally, other causes such as congenital anomalies or tumors may be responsible. Thrombotic EHPVO in the absence of inflammatory diseases such as pancreatitis is essentially an Asian disease, being reported primarily from the Indian subcontinent and Japan. In these areas, it is the commonest cause of portal hypertension in children and young adults [1]. In comparison, in Europe or North America, EHPVO is very rare, being observed occasionally among elderly persons [1]. Similarly, HVOTO in Asian patients appears to be a different disease from that in the West because of an earlier age of onset, a subacute or insidious onset, and a longer survival [2].

In Western countries, “Philadelphia chromosome”-negative myeloproliferative disorders, namely polycythemia vera, essential thrombocythemia, and myelofibrosis, are a major cause for both HVOTO and EHPVO, being responsible for 33 % to 78 % and 23 % to 48 % of cases with these conditions, respectively [3–6]. These myeloproliferative disorders are related to a defect in multipotent hematopoietic stem cells which leads to an increased production of mature blood cells. The increased and autonomous production of these cells, important for the diagnosis of myeloproliferative disorders, is however difficult to establish in patients with HVOTO or EHPVO because of the frequent presence of hemodilution, occult bleeding, and hypersplenism [7–10].

In 2005, an acquired mutation in the Janus kinase 2 (JAK2) gene located on the short arm of chromosome 9 was identified in patients with polycythemia vera, essential thrombocythemia, and idiopathic myelofibrosis [11–15]. In this mutation, a guanidine-to-thymidine change at nucleotide 1849 in exon 12 leads to substitution of valine by phenylalanine at amino acid position 617 (V617F) of the JAK2 protein. This change, located in the JH2 pseudokinase domain, is associated with a gain of function and makes the JAK2 protein constitutively active even in the absence of erythropoietin, leading to an increased proliferation of hematopoietic stem cells. Several studies from the West have reported the presence of this mutation in 20 % to 45 % and 35 % of patients with HVOTO and EHPVO, respectively [16–18].

However, data on the role of myeloproliferative disorders and of this mutation in Indian patients with HVOTO and EHPVO are limited. In view of the onset at young age and relatively benign long-term course of these conditions in this geographical area, it is possible that myeloproliferative disorders contribute much less to the causation of these conditions in India. To clarify this issue, we studied the frequency of JAK2 V617F mutation in Indian patients with HVOTO and EHPVO.

Methods

Subjects

Patients with HVOTO or EHPVO attending the outpatient or inpatient services of our department between January and October 2011 were invited to participate in the study. The study protocol was approved by the Ethics Committee of our institution, and all subjects provided an informed consent.

All the patients had clinical and/or endoscopic evidence of portal hypertension. Diagnosis of HVOTO was based on radiological evidence of obstruction of the main hepatic veins, inferior vena cava, or both. Diagnosis of EHPVO was based on evidence of obstruction of the extrahepatic portal vein, with or without its tributaries, along with demonstration of a portal cavernoma either at Doppler ultrasound or at magnetic resonance portovenography. Patients with identifiable causes of splanchnic venous thrombosis (i.e. liver tumors, pancreatitis, liver cirrhosis, previous laparotomy, recent pregnancy, or use of oral contraceptives) were excluded.

Detection of JAK2 V617F mutation

From each patient, 2 mL of peripheral blood was collected in EDTA. DNA was extracted using a standard spin column technique (QIAmp DNA Mini Kit; Qiagen, Valencia, CA, USA) as per the manufacturer’s instructions. Two separate DNA amplification methods were used for detection of JAK2 V617F mutation in DNA from each patient. DNA from one healthy person and from two patients with polycythemia rubra vera previously shown to have the V617F mutation were used as negative and positive controls, respectively.

Method A

This PCR-based method used three primers (two forward and one common reverse; Table 1) [12]. The common forward and common reverse primers amplify both wild-type and mutant DNA to produce a 364-base pair (bp) long amplicon. The third primer binds only to mutant DNA and amplifies an additional PCR product of 203 bp along with the common reverse primer. In the absence of JAK2 V617F mutation, this shorter amplicon is not generated. The generation of longer amplicon serves as an internal control to confirm that the amplification worked (Fig. 1).

Table 1 Nucleotide sequences of primers used
Full size table
Fig. 1
figure 1

Principle of method A. The primers 1 and 2 are common forward and reverse primers, respectively, that can bind to both wild-type and mutant DNA. Primer 3 is a forward primer that is specific for the mutant DNA sequence. Wild-type DNA produces only one amplicon (364 bp), whereas the presence of mutant DNA leads to production of two amplification products (364 and 203 bp)

Full size image

Method B

This method used four primers: two forward and two reverse (Table 1) [19]. The outer primers bind to both wild-type and mutant DNA generating a 463-bp product, whose production serves as an internal reaction control. The two internal primers, one forward and one reverse, are located in the middle and are specific for wild-type and mutant DNA, respectively. The former generates a 229-bp PCR product with the wild-type DNA but not with the mutant DNA, and the latter generates a 279 bp long with the mutant DNA and not with the wild-type DNA. On electrophoresis of PCR product, a mixture of wild-type and mutant DNA would be expected to provide three bands (i.e. 463, 229, and 279 bp), whereas pure wild-type DNA would provide only two bands (463 and 229 bp; Fig. 2).

Fig. 2
figure 2

Principle of method B. The outer primers 1 and 2 are common forward and reverse primers, respectively, that can bind to both wild-type and mutant DNA. Primer 3 is a forward primer that is specific for the wild-type DNA sequence and in association with the common reverse primer produces a 229-bp product. Primer 4 is a reverse primer that is specific for the mutant DNA sequence and in association with the common forward primer produces a 279-bp product. Thus, wild-type DNA produces only two amplification product of 463 and 229-bp length, whereas the mutant DNA produces two amplification products of 463 and 279-bp length

Full size image

Confirmation of mutation using DNA sequencing

For any specimen showing JAK2 V617F mutation using either of the above methods, the PCR products generated using external primers for method B above were sequenced using the dideoxynucleotide chain termination method (ABI 3130 genetic analyzer; Applied Biosystems, Foster City, CA, USA).

Results

A total of 90 patients, including 40 with HVOTO and 50 with EHPVO, were studied. Clinical and laboratory profiles of these patients are shown in Table 2. Among patients with HVOTO, venous obstruction was restricted to one or more hepatic veins in 23 (57.5 %), involved the hepatic veins as well as the IVC in 15 (37.5 %) patients, and only the IVC in 2 (5 %) patients. In one patient with combined hepatic and IVC venous obstruction, the IVC occlusion was caused by a membranous lesion.

Table 2 Clinical and laboratory test profiles of patients with hepatic venous outflow tract obstruction and extrahepatic portal venous obstruction included in the study
Full size table

Using both methods, 1 of the 50 patients with EHPVO and none of 40 patients with HVOTO tested positive for JAK2 V617F mutation (Figs. 3 and 4). DNA sequencing confirmed the existence of a guanidine-to-thymidine change at nucleotide 1849 in exon 12 of the JAK2 gene in this patient.

Fig. 3
figure 3

Gel electrophoresis for method A. The 364-bp band is expected with both wild-type and mutant DNA. In contrast, the 203-bp product is expected only with mutant DNA, as shown in lane 1 (positive control). None of the patient specimens shown in this picture yielded the 203-bp product, indicating the absence of JAK2 V617F mutation in these patients

Full size image
Fig. 4
figure 4

Gel electrophoresis for method B. The 463-bp band is expected with both wild-type and mutant DNA; the 229-bp band is expected only with wild-type DNA, and the 279-bp band is expected only with mutant DNA. None of the patient specimens included in this picture yielded the 279-bp product, suggesting the absence of mutation. Positive control sample yielded only 279-bp product (but no 229-bp product), indicating homozygous JAK2 V617F mutation

Full size image

The patient with JAK2 V617F mutation was a 24-year-old man with non-specific abdominal pain along with dragging sensation in the left upper abdomen for 3-year duration. His investigations revealed hemoglobin 10.9 g/dL, total leukocyte count 9000/μL, platelet count 251,000/μL, bilirubin 3.7 mg/dL, alanine aminotransferase 25 IU/mL (normal <40), aspartate aminotransferase 53 IU/mL (normal <40), alkaline phosphatase 129 IU/mL (normal <150), and prothrombin time 15.2 s (control 11.8 s). Upper gastrointestinal endoscopy revealed grade II esophageal varices. HBsAg and anti-HCV tested negative. Bone marrow biopsy revealed a hypercellular bone marrow (cellularity 80 % to 90 %), with moderately increased and pleomorphic megakaryocytes, normal maturation of erythroid cells, and increased eosinophilic cells, suggesting the presence of myeloproliferative disorder. The number of blasts, plasma cells, and lymphocytes were within normal limits.

Discussion

In this study that employed two separate techniques for detection of V617F mutation in the JAK2 gene, only 1 of 50 patients with EHPVO and none of 40 patients with HVOTO was found to have this mutation. These rates are much lower than those reported in studies from Western countries.

Some previous studies from India have looked for the presence of one of the myeloproliferative disorders as underlying cause for HVOTO. The first such Indian study used bone marrow culture technique and looked for the formation of endogenous erythrocyte colonies, and 21 (69 %) of the 32 patients with HVOTO studied tested positive [20]. The presence of such colonies was taken as evidence of a latent myeloproliferative disorder, which thus appeared quite frequent. In addition, individual case reports of HVOTO associated with myeloproliferative syndrome and JAK2 V617F mutation have been reported. [21, 22]. However, in two recent studies, myeloproliferative disorder was detected in only 8.5 % and 8.8 % of Indian patients with HVOTO [23, 24]. In one of these studies, JAK2 V617F mutation was also looked for, and 12 of 137 HVOTO patients were found to be positive for this mutation. These data suggest that JAK2 V617F mutation and hence myeloproliferative disorders are not a predominant cause of HVOTO among Indian patients. This is in contrast to the Western reports in which myeloproliferative disorders accounted for more than 50 % of all cases with HVOTO and a known cause. Our study is in agreement with the previous Indian data. However, in one other study, JAK2 V617F mutation was observed in 6 of 32 patients with portal vein thrombosis and in 4 of 20 patients with hepatic vein thrombosis [25]. But, this report is from a hematology unit and hence possibly preferentially included patients with myeloproliferative disease. Similarly, another Indian report showed JAK2 V617F mutation in 40 % of Indian patients with HVOTO and 15 % of those with EHPVO [26]. However, this study included a large number of patients who were already known to have a myeloproliferative condition or acute venous thrombosis. The results of these studies cannot be compared to those of our study, since the latter included only patients with idiopathic HVOTO or EHPVO.

The clinical presentation of HVOTO in Asian patients is also different from that of patients with this disease in Western countries. Whereas in Western countries, the disease is associated with a female preponderance and relative acute presentation, Asian patients with this disease are often characterized by a long duration of symptoms and signs and a higher prevalence of liver cirrhosis. It has been suggested that these geographical differences are related to differences in the location of the venous blockage, with pure hepatic vein block accounting for 62 % of HVOTO cases from Western countries [27], but only 6 % of Japan cases and 24 % of those in India; in the latter regions, obstruction of the IVC is more common [2, 28]. However, this explanation may not be true since in recent years, an increasing proportion of Asian patients with HVOTO have been reported to have only hepatic vein block. For instance, in our study, more than half of the patients with HVOTO had obstruction limited to hepatic veins. Similar predominance of hepatic vein involvement has been reported in another Indian study, wherein obstruction was present in the hepatic vein in 59 % of patients [23]. This is contrary to most of previous studies from India where combined hepatic plus IVC obstruction was the predominant site, in frequency of 56 % to 57.5 % [23, 29, 30]. Even in comparing those Asian HVOTO patients who have only hepatic vein block, differences from Western patients with HVOTO are apparent.

EHPVO is a condition that predominantly occurs in children and adolescents in Asian countries. The precise reason for the development of EHPVO in children has not been identified. Data that are available on JAK2 mutation in Asian patients with EHPVO are limited. We found that only 1 (2 %) of our 50 patients had the JAK2 mutation. A previous Indian study had found JAK2 V617F mutation in only 4 of 76 (5 %) patients with portal vein thrombosis [24]. In contrast, most of the Western studies have shown the presence of JAK2 mutation in nearly one third of cases with portal venous thrombosis. Interestingly, a study from UK showed that none of 30 EHPVO patients studied there had JAK2 V617F mutation [31]; it was found that a large proportion of these patients were of Asian origin.

Apart from myeloproliferative disorders, several other inherited and acquired prothrombotic disorders have been found to be associated with splanchnic venous thrombosis. Thus, HVOTO is a multifactorial disease, which can be caused by several prothrombotic disorders, occurring alone or simultaneously [3, 32]. Among various inherited prothrombotic conditions, factor V Leiden mutation and protein C and S deficiency have been shown to associate with HVOTO in around 25 %, 12 %, and 7 %, respectively, in Indian studies [33, 34], similar to data from the West [3, 35, 36]. However, a study from our center had failed to find an association between HVOTO and factor V Leiden and prothrombin gene G20210A mutations [29].

Similarly, no cause can be found for most of the cases with EHPVO. Various hypotheses have included omphalitis, neonatal umbilical sepsis and umbilical vein cannulation [37], congenital defects of portal vein [38], and inherited prothrombotic disorders [33, 39]. However, no cause can be found in a large majority of patients with EHPVO. Previous studies have failed to show a strong association of inherited thrombotic disorders, factor V Leiden mutation, prothrombin gene G20210A mutation, and protein C and S deficiency with EHPVO among Indian children [40–42]. Our data show that even the common JAK2 mutation is unlikely to be responsible for this condition in most of the patients. This suggests the need for further etiological studies to identify the cause of HVOTO and EHPVO in Asian patients.

In conclusion, our data indicate that JAK2 V617F gain-of-function mutation, which is a marker of myeloproliferative disorders, was noted as a major etiological factor for HVOTO and EHPVO in our patients. Further studies are needed to identify the cause(s) underlying these diseases related to venous thrombosis, which appear to be particularly common as causes of portal hypertension in India and neighboring countries.