Relationship Between Ebola Virus Real-Time Quantitative Polymerase Chain Reaction–Based Threshold Cycle Value and Virus Isolation From Human Plasma

Abstract

We performed a longitudinal analysis of plasma samples obtained from 4 patients with Ebola virus (EBOV) disease (EVD) to determine the relationship between the real-time quantitative reverse transcriptase polymerase chain reaction (qRT-PCR)–based threshold cycle (Ct) value and the presence of infectious EBOV. EBOV was not isolated from plasma samples with a Ct value of >35.5 or >12 days after onset of symptoms. EBOV was not isolated from plasma samples in which anti–EBOV nucleoprotein immunoglobulin G was detected. These data demonstrate the utility of interpreting qRT-PCR results in the context of the course of EBOV infection and associated serological responses for patient-management decisions.

Ebola virus (EBOV) is a single-stranded, negative-sense RNA virus that can cause disease in humans, with a case-fatality rate as high as 90%. Until the current West African outbreak, EBOV had only caused sporadic outbreaks of EBOV disease (EVD) in Central Africa. EBOV can be shed in a wide variety of bodily fluids, and direct contact with bodily fluids is considered to be the major risk factor for infection [1].

Testing was historically performed using an antigen-capture enzyme-linked immunosorbent assay (ELISA). An agarose gel electrophoresis–based polymerase chain reaction (PCR) for detection of filovirus in samples from patients who acquired EVD during the outbreak in Kikwit, Democratic Republic of the Congo, was used in 1995 [2]. The first use of nested PCR in an EVD response by the Centers for Disease Control and Prevention (CDC) was in 2000, in Uganda, for detection of Sudan virus [3]. Real-time quantitative reverse transcriptase PCR (qRT-PCR) was developed and optimized by retrospective analysis of the Sudan virus samples and first used in the field for filovirus diagnostic testing in Angola 2005 [4, 5]. The qRT-PCR assay currently in use by the CDC assesses EBOV load by primers specific for the viral nucleocapsid gene. Importantly, interpretations of these data have direct implications for primary patient care and the public health response. Discharging patients with EVD from EVD treatment units (ETUs) as soon as it is safe to do so from both a patient and public health perspective is critical to outbreak response. The ongoing outbreak has been complicated by the large number of cases and lack of space within care facilities to manage and treat patients. The decision to release patients from ETUs is based on resolution of clinical symptoms, in addition to qRT-PCR evidence that they have cleared EBOV from blood. Current protocols for release of patients from ETUs vary but in general require negative results of 2–3 sequential PCR tests of blood specimens, defined as 40 cycles with undetectable viral RNA.

Previously, virus isolation was only attempted from samples with high levels of EBOV RNA, as assessed by the threshold cycle (Ct) value. The qRT-PCR-based Ct value at which virus isolation is no longer possible was estimated to be 35–39, based on anecdotal evidence from field isolates and findings from experimental animal studies of related filoviruses that provided the only longitudinal data available [6, 7]. Herein, we report a longitudinal study of plasma samples obtained from patients with EVD that were analyzed for EBOV RNA load, underwent serological analysis, and subsequently were subjected to viral isolation testing, to describe the relationship between qRT-PCR-based viral load assessment in plasma and the ability to isolate infectious EBOV.

METHODS

Sample Collection

Samples were obtained 2–23 days after onset of patient-reported symptoms from 4 patients with EVD during care at Emory University. All patients provided informed consent (CDC institutional review board [IRB] protocol 6643; Emory IRB protocol 00076700). Plasma was obtained from whole-blood samples (collected in tubes containing ethylenediaminetetraacetic acid or citrate) that were collected as part of routine patient management.

RNA Extraction and qRT-PCR

Total RNA was purified from plasma samples, using the MagMAX-96 Viral RNA Isolation Kit on the Invitrogen BeadRetriever system. qRT-PCR was performed with the EBOV nucleoprotein 1 (NP-1) assay [8], using the SuperScript III Platinum One-Step qRT-PCR Kit (Life Technologies) on the Applied Biosystems 7500 Real-Time PCR System.

Virus Isolation and IFA

Vero-E6 cells grown to confluence were inoculated with patient plasma (100 μL of inoculum per 25-cm² flask) obtained at various time points during EBOV infection. Flasks were monitored for CPE, and supernatants were screened by RNA isolation and qRT-PCR 1 and 2 weeks after inoculation. Two weeks after inoculation, supernatants were collected to assay live virus. Supernatants (100 μL) were inoculated onto Vero-E6 cells seeded to confluence on glass coverslips. One week later, cells were fixed in formalin and subjected to γ irradiation. Cells were permeabilized with 0.1% Triton-X, washed with phosphate-buffered saline (PBS), blocked with 1% bovine serum albumin–PBS, and stained with EBOV rabbit polyclonal antibody (CDC reference stock 703371) at a dilution of 1:500. Secondary staining was performed with anti-rabbit-FITC at a dilution of 1:1000 (Life Technologies), and cells were visualized by immunofluorescence microscopy.

ELISAs

NP-specific ELISAs were performed as previously described by McElroy et al [9]. ELISA data included in this has article have been replicated from the article by McElroy et al to allow for comparison with respective Ct values and isolation attempts.

RESULTS

A total of 28 plasma samples from 4 patients with EVD (EVD2, EVD5, EVD9, and EVD15) were tested by qRT-PCR and subsequently used for virus isolation attempts. During the course of treatment, 2 of 4 patients received ZMapp, an experimental biopharmaceutical drug comprising 3 chimeric monoclonal antibodies that is under development as a treatment for EVD. EVD2 received 1 unit of convalescent whole blood on day 7 and ZMapp on days 9, 12, and 15 after onset of symptoms. EVD5 received ZMapp on days 10, 13, and 16 after onset of symptoms [10]. EVD9 was treated with convalescent plasma (on days 8, 9, 11, 12, 14, and 15) and TKM-Ebola (on days 3–8), an RNAi therapeutic produced by Tekmira Pharmaceuticals, but did not receive ZMapp [11]. EVD15 received brincidofovir (on days 1 and 4) and convalescent plasma (on days 2 and 3).

EBOV was isolated from plasma samples from 3 of 4 patients at various times after onset of symptoms (Table 1 and Figure 1). No EBOV isolates were obtained from patient samples collected ≥12 days after onset of symptoms or from plasma samples with Ct values of ≥36. Immunoglobulin M (IgM) and immunoglobulin G (IgG) were detected at various times after onset of symptoms in all 4 patients with EVD (Table 1 and Figure 1). EBOV was isolated from samples with detectable NP-specific IgM but not from any samples with detectable NP-specific IgG.

Table 1.

Threshold Cycle (Ct) Values, Serological Findings, and Corresponding Virus Isolation Findings During the Course of Illness

Patient, Day After Symptom Onset	Ct Value	IgM Titer	IgG Titer	Virus Isolation Finding
EVD2
12a	26.0	1:72 900	Negative	Positive
14	30.8	1:24 300	1:300	Negative
15a	32.8	1:24 300	1:900	Negative
16	34.8	1:72 900	1:24 300	Negative
17	33.9	1:72 900	1:72 900	Negative
EVD5
16a	36.1	1:24 300	1:24 300	Negative
17	>40	1:24 300	1:24 300	Negative
18	36.7	1:72 900	1:72 900	Negative
20	37.5	1:24 300	1:218 700	Negative
23	37.4	1:72 900	1:72 900	Negative
EVD9
5b	23.2	Negative	Negative	Positive
6b	21.5	Negative	Negative	Positive
7b	20.6	Negative	Negative	Positive
8b,c	19.8	1:2700	Negative	Positive
9c	24.0	1:2700	Negative	Positive
10	23.0	1:2700	Negative	Positive
11c	22.4	1:2700	Negative	Negative
12c	23.6	1:2700	Negative	Negative
13	26.5	1:24 300	Negative	Negative
14c	30.7	1:2700	Negative	Negative
15c	34.0	1:2700	1:300	Negative
16	35.0	1:2700	1:300	Negative
17	35.1	1:2700	1:2700	Negative
18	35.8	1:2700	1:8100	Negative
EVD15
2c	35.4	Negative	Negative	Positive
3c	36.8	1:100	1:2700	Negative
4	37.1	1:900	1:8100	Negative
6	>40	1:8100	1:24 300	Negative

Abbreviations: EVD, EBOV disease; IgG, immunoglobulin G; IgM, immunoglobulin M.

^a Received ZMapp.

^b Received TKM-Ebola.

^c Received convalescent plasma.

Figure 1.

Time line of real-time quantitative reverse transcriptase polymerase chain reaction (qRT-PCR)–based threshold cycle (Ct) values and antibody titers in plasma. Median Ct values are denoted by the solid line, and median immunoglobulin G (IgG) and immunoglobulin M (IgM) titers are denoted by dashed lines.

DISCUSSION

qRT-PCR is now the diagnostic method of choice for both field diagnostic tests and reference laboratory assessment of EBOV infection. Prior to this report, the relationship between the EBOV qRT-PCR Ct value and the ability to isolate virus had not been investigated in detail. In our study, use of EBOV NP1 qRT-PCR in longitudinal analysis of plasma samples from 4 patients with EVD revealed that Ct values of >35.5 were not associated with detection of infectious EBOV. In addition, samples obtained from patients later in the course of EVD, 12 days after onset of symptoms, even those with Ct values of <31, were not associated with detection of infectious EBOV.

Importantly, the effect of time after onset of symptoms on the presence of infectious EBOV was associated with development of anti-EBOV IgM and IgG immune responses, suggesting that the immune response, rather than the time after symptom onset per se, was responsible for the inability to isolate infectious viruses from these patients. In patients who survive, antigen has been found to decline to undetectable levels by day 14 or 15 after symptom onset, as measured by the classical antigen assay [12]. The timing during the course of disease is a reflection of the patient's immune response to EBOV infection, and IgG levels appear to be a key correlate of protection. Our data suggest that a virus-specific IgM response does not necessarily indicate clearance of infectious EBOV from plasma. In contrast, the presence of virus-specific IgG, especially early in the course of disease, correlates with clearance of infectious virus. ZMapp IgG treatment did not contribute to IgG levels reported herein, as an EBOV NP antigen ELISA was performed and ZMapp is specific to glycoprotein (GP). For EVD9, virus isolation ceased prior to the development of NP-specific IgG. However, it is not clear how convalescent plasma treatment may have influenced the course of disease, especially if there were GP-specific antibodies present. In addition, in contrast to the other patients, EVD9 had a prolonged IgM response prior to IgG detection. While the presence of anti-EBOV IgM does not appear to preclude virus isolation from plasma, it may contribute to viral clearance in certain cases. Our findings are supported by a recent EVD case study that, similarly, reported virus isolation from plasma up to 14 days after onset of symptoms, prior to development of anti-EBOV IgG [13], suggesting neutralization of virus by anti-EBOV IgG. Furthermore, investigation of vaccine candidates and therapeutics indicates a key role for IgG: levels of EBOV GP–specific IgG antibody measured immediately before challenge correlated with survival in animal models [14, 15]. In addition, polyclonal IgG from rhesus macaques that survived an EBOV challenge was shown to protect naive rhesus macaques when administered up to 48 hours after infection [16].

A similar relationship between the course of disease and the presence of infectious virus in plasma may be seen in other filovirus infections. An isolate was obtained from a Marburg virus–infected patient early in the course of disease, with a corresponding Ct value of 32. However, later samples obtained from the same patient on days 10 and 20 after onset of symptoms were both tested negative by virus isolation assays, with associated Ct values of 34 and undetectable, respectively [6]. With regard to filovirus infection, an important caveat in translating these findings into a public health setting is that these data are based exclusively on plasma samples. Additional investigations performed on other bodily fluids to confirm the absence of infectious particles would further support modified criteria for release of patients from ETUs.

Our study could have implications for EVD management and discharge consideration from ETUs because these findings provide evidence supporting the discharge of patients, contingent on clinical status, prior to negative results of qRT-PCR. In Liberia, EVD survivors are currently estimated to stay a mean of 11.8 days in the ETU [17]. While decreased length of stay would increase capacity, there are other ongoing issues that must be addressed in conjunction with appropriate release of patients. Patients who do not survive have an estimated length of stay of 4.2 days [17]. While a more liberal discharge policy would clearly not affect the patients who do not survive, it would provide a significant beneficial influence on patients who are asymptomatic and awaiting clearance by qRT-PCR data to be discharged. Faster discharge of recovered patients and more-rapid reunion of families would go far to build patient and community trust.

On the basis of the limited data reported here, alternative discharge criteria could be considered for patients who have a EBOV NP1 assay Ct value of ≥36 and are asymptomatic. Additionally, future studies should address the use of serological assay findings as a possible indicator that it is safe from a public health standpoint to discharge asymptomatic patients with EVD.