First Isolation of West Nile Virus from a Patient With Encephalitis in the United States

Cinnia Huang, Brett Slater, Robert Rudd, Nandakishore Parchuri, Rene Hull, Michelle Dupuis, Alexander Hindenburg


Emerging Infectious Diseases. 2002;8(12) 

In This Article


CSF was simultaneously examined for a panel of 11 viruses by RT-PCR/PCR as described in Material and Methods. No amplification product was observed for any other virus in the PCR battery except WNV. Two primer pairs (CU9093/CL9279 and D87F/D156R) in the NS5 region[5,6] of the WNV genome were used in the initial screening by standard RT-PCR ( Table 1 ). PCR products of appropriate size for both primer pairs were obtained (Figure 1). RT-PCR was independently repeated on another aliquot of CSF by using four primer pairs located in various genomic regions of WNV. PCR bands with the expected sizes were present in all four reactions (data not shown). The identity of the PCR bands was confirmed by sequence data obtained directly from the PCR amplicon. The diagnosis of WNV infection was based on the detection and sequence of the WNV genome in CSF. Although the serologic test (IgM capture enzyme-linked immunosorbent assay [ELISA]) on the CSF sample was negative for WNV, according to the interpretations set forth by CDC, the detection of viral genome sequence in CSF meets the definition of a confirmed case.

RT-PCR detection of West Nile virus RNA in cerebrospinal fluid. Lanes 1 and 3: negative controls; lanes 2 and 4: cerebrospinal fluid; lane M: 50-bp DNA ladder. Primer pairs used: lanes 1 and 2: CU9093/CL9279; lanes 3 and 4: D87F/D156R.

To follow up this case, three whole-blood samples and one serum sample were examined by both standard RT-PCR (primers in NS5 region) and TaqMan (primers in ENV region) assays; the results are summarized in Table 2 . The highest RNA copy number, 2.5 x 106 copies/mL, was found in the blood sample that was collected 3 days after the patient's neurologic symptoms appeared. WNV genome was also detected in a serum sample collected on day 19 after onset of symptoms. Serologic tests (IgM capture ELISA and IgG ELISA) on serum for Eastern equine encephalitis virus, LACV, POWV, SLEV, and WNV were all negative.

Laboratory studies of the CSF indicated the following values: leukocyte (WBC) count 8/mm3 with 66% neutrophils, 4% lymphocytes, 4% atypical lymphocytes, and 26% monocytes; erythrocyte count 0; glucose level 76 mg/dL; total protein level 55 mg/dL, and lactate dehydrogenase level 35 IU/L. Figure 2 presents fever curve, WBC curve, and viremia data. The serum immunoglobulins at the time of infection with WNV were IgG 492 mg/dL (normal level [nl]) 700-1,500 mg/dL), IgA 86 mg/dL (nl 65-450 mg/dL), and IgM 80 mg/dL (nl 45-230 mg/dL).

West Nile virus copy numbers in clinical samples and clinical indices. WBC, leukocytes. Detailed sample information is listed in Table 2; day 1 is date the patient was hospitalized, 9/18/2001.

The attempt to isolate WNV was carried out in a biosafety level 3 laboratory not routinely used for arbovirus work and located in a building separate from the facility where PCR testing was conducted. This procedure had the dual advantage of minimizing the possibility that any virus recovered originated from a source other than the human specimen and also ensuring that any isolate obtained would not lead to future spurious PCR results. A blood sample collected on September 24, 2001, and a CSF specimen collected on September 26, 2001, were chosen for recovery of the virus because of the presence of high-copy-number viral RNA. On day 6 postinfection, CPE was observed in the tubes inoculated with 0.01 mL and 0.05 mL of blood, but not in the tube inoculated with 0.1 mL of CSF. WNV was confirmed in the second-passage cell cultures by IFA by using WNV-specific monoclonal antibody H5-46.

Since only a very small amount of CSF remained after the diagnostic work-up, we carefully designed a protocol to generate PCR bands with sizes ranging from 500 to 1000 bp. This protocol allowed the complete genome sequence of the virus in the CSF to be determined by sequencing overlapping PCR bands (GenBank accession no. AF533540). We used a similar protocol to obtain the genome sequence of the isolate adapted from cell culture and found that the sequence data from the virus in the CSF and from the WNV isolate were identical. A 1648-bp fragment encoding the PreM, M, and part of the 5'-E gene was used for phylogenetic studies (Figure 3). The analysis showed that the sequence data from this case are similar to the sequence data obtained from human[7] (GenBank accession no. AF202541), horse[8] GenBank accession no. AF260967), and bird[9] (GenBank accession no. AF196835) WNV isolates in New York in 1999.

Phylogenetic relationships among West Nile virus strains. Sequence data from the present case are shown in italics. The tree is based on the 1,648-bp fragment encoding the preM, M, and part of the 5¢-E gene. Numbers at the nodes are bootstrap confidence estimates based on 1,000 replicates.

This case is important for several reasons. It represents the first instance in which WNV was recovered from a person in the United States. It is also the first time that the entire genomic sequence of WNV has been obtained from CSF from a human case-patient. The sequence data from the virus directly detected in the CSF and from the WNV isolate from cell cultures are identical. Hindiyeh et al.[10] reported the isolation of WNV from the blood of viremic patients who were not immunocompromised and who seroconverted later. In contrast, the patient in this case was elderly and was undergoing treatment for lymphoma. She was unable to mount an immune response as shown by the fact that the results of serologic tests on both CSF and serum specimens were negative.

To our knowledge, all previous attempts to recover WNV from human patients associated with the North American outbreak have been unsuccessful. The ability to recover an infectious isolate in this report may have been contingent on the fact that the patient was immunologically impaired. She had lymphoma and had been undergoing treatment with CHOP plus rituximab for 2 months at the time the WNV infection developed. The relative role of immunosuppression caused by the lymphoma itself compared with the immunosuppression due to treatment of the lymphoma is unclear. At the time of the WNV infection, the patient's serum IgG levels were moderately suppressed. This was most likely the result of lymphoma, because a reduction in immunoglobulins, secondary to impaired B-cell function from rituximab, usually occurs after 3 months of therapy[11]. In a previous randomized study, elderly patients receiving CHOP chemotherapy and rituximab had increased their overall survival and had not experienced an increase in toxic clinical effects compared with effects from CHOP treatment alone[12]. However, rituximab used as a single agent has been reported to lead to excessive bacterial and viral infections, including respiratory tract infections and herpes[13]. Rituximab is also implicated as a risk factor for unusual viral infections when used as an immunotherapy agent in the peritransplant period of autologous stem cell transplant in non-Hodgkin lymphoma patients[14]. Another consideration is the fact that the patient was neutropenic. The relationship between neutrophil function and the severity of WNV infection is unknown. The virus may be cleared by neutrophils, and the severity of the viral infection may have been due to the fact that the patient was neutropenic at the time of the acute infection. What lends credence to this hypothesis is the observation that the highest viral titer as determined by PCR coincided with the recovery of the WBC count. Following the resolution of the myelosuppression, the RNA copy number of the WNV in blood samples declined rapidly (from 1.1 x 106 to 5.4 x 104 copies/mL).

In summary, this report is the first of WN encephalopathy in an immunocompromised patient undergoing treatment for lymphoma. The patient's serologic tests remained negative, and the diagnosis was made by RT-PCR from both CSF and peripheral blood, and by in vitro cultivation of the virus from blood. WNV infection should therefore be considered in the differential diagnosis of patients with lymphoma who exhibit encephalopathy, even if serologic tests are negative for WNV. Extra care should be taken to prevent patients with lymphoma, especially those undergoing treatment, from being exposed to WNV.