Evidence that Ecotropic Murine Leukemia Virus Contamination in TZM-bl Cells Does Not Affect the Outcome of Neutralizing Antibody Assays with Human Immunodeficiency Virus Type 1

The TZM-bl cell line that is commonly used to assess neutralizing antibodies against human immunodeficiency virus type 1 (HIV-1) was recently reported to be contaminated with an ecotropic murine leukemia virus (MLV) (Y. Takeuchi, M. O. McClure, and M. Pizzato, J. Virol. 82:12585-12588, 2008), raising questions about the validity of results obtained with this cell line. Here we confirm this observation and show that HIV-1 neutralization assays performed with a variety of serologic reagents in a similar cell line that does not harbor MLV yield results that are equivalent to those obtained in TZM-bl cells. We conclude that MLV contamination has no measurable effect on HIV-1 neutralization when TZM-bl cells are used as targets for infection.It was recently reported that TZM-bl cells, which are commonly used to assess neutralizing antibodies (Abs) against human immunodeficiency virus type 1 (HIV-1), are contaminated with an ecotropic murine leukemia virus (MLV) (22). TZM-bl (also called JC.53bl-13) is a HeLa cell derivative that was engineered by amphotropic retroviral transduction to express CD4 and CCR5 (17) and was further engineered with an HIV-1-based vector to contain Tat-responsive reporter genes for firefly luciferase (Luc) and Escherichia coli β-galactosidase (24). These engineered features made TZM-bl cells highly susceptible to HIV-1 infection in a readily quantifiable assay for neutralizing Abs. Many published studies used this cell line for assessments of HIV-1 neutralization; these include several recent reports describing the magnitude, breadth, and epitope specificity of the neutralizing Ab response in infected individuals (14, 18-20), neutralization escape (25), and the neutralization phenotype of transmitted/founder viruses (10). TZM-bl cells are also gaining popularity for assessments of vaccine-elicited neutralizing Ab responses (13). The validity of these and other published results, together with a rationale for the continued use of TZM-bl cells in assessing neutralizing Abs against HIV-1, are very dependent on establishing to what extent, if any, MLV contamination affects the outcome of the assay.It was suggested that ecotropic MLV entered TZM-bl cells via the progenitor JC.53 cell line as an amphotropic MLV pseudotype (22). In this regard, JC.53 cells were constructed from HeLa cells in two stages by using ping-pong technology to amplify the pSFF vector derived from the replication-defective and highly truncated Friend spleen focus-forming virus (3). When used with this vector, this procedure has previously resulted in stable vector expression (17) without formation of replication-competent MLV recombinants (8, 11). A panel of HeLa-CD4 clones was made that express different amounts of CD4 and where the high-expression HI-J clone was used to make a derivative panel of clones (termed JC), including JC.53, that expressed diverse levels of CCR5 (9, 16, 17). In addition, the HeLa-CD4 clone HI-R that expressed low levels of CD4 was used to make another panel of CCR5-expressing clones (termed RC). To investigate this newly reported issue, cell extracts from these clonal panels and from TZM-bl cells were analyzed for MLV Gag antigens by Western immunoblotting. Representative data, as shown in Fig. ​Fig.1A,1A, confirm that JC.53 and TZM-bl cells express MLV Gag antigens, whereas the progenitor HI-J clone of HeLa-CD4 cells and many but not all of the other HeLa-CD4/CCR5 clones in the JC panel lack MLV antigens.Open in a separate windowFIG. 1.Characterization of HeLa clones for MLV Gag expression, HIV-1 susceptibility, and cell surface expression of HIV-1 fusion receptors. (A) MLV Gag antigen expression in HeLa cells and derivative clones expressing CD4 or CD4 and CCR5. Cell lysates were prepared from the cell clones and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting with Abs to MLV Gag antigens (upper blot). The lysates were also probed with anti-tubulin antibodies (lower blot). Lane 1, HeLa cells; lanes 2 and 3, HeLa CD4 clones HI-R and HI-J, respectively; lanes 4, 5, and 6, HeLa-CD4/CCR5 clones JC.10, JC.48, and JC.53, respectively; lane 7, TZM-bl cells; lane 8, psi-2 packaging cells positive for MLV Gag. (B) HIV-1 infectivity on the HeLa-CD4/CCR5 JC panel. Target cells were infected with HIV-1 isolate JRCSF that had been produced from clone JC.53 cells (black) or with JRCSF produced from transfected HEK293T cells (red). The target cells were also infected with the JR-FL isolate produced from peripheral blood mononuclear cells (PBMC; green). The HeLa-CD4/CCR5 target cells had a CCR5 expression range of 2 × 10³ (clone JC.10) to 1.3 × 10⁵ (clones JC.53 and TZM-bl) CCR5 molecules/cell. Each set of three data points at a given CCR5 expression level represents a single HeLa-CD4/CCR5 JC clone. None of the HIV-1 isolates was able to infect HeLa-CD4 cells lacking CCR5. The blue asterisks indicate clones that are negative for MLV Gag proteins. Clones JC.48 (used for subsequent infection and neutralization assays) and JC.53 (progenitor of TZM-bl cells) are specifically labeled. (C) Surface expression of CD4, CCR5, and CXCR4 on TZM-bl and JC.48 cells was assessed by flow cytometry using the same stocks of cells that were used in infection and neutralization assays in Fig. ​Fig.2.2. Surface staining was performed with phycoerythrin-conjugated mouse monoclonal Abs to CD4, CCR5 (CD195), and CXCR4 (CD184). Background staining was performed with isotype-matched control Abs. All Abs for flow cytometry were purchased from BD Biosciences Pharmingen (San Diego, CA). Results are shown as the mean fluorescence intensity (MFI) of positive cells. Most cells (>90%) stained positive in each case.Initial studies of HI-R cells and other clonal panels that were made using these methods also suggested a lack of MLV antigens (data not shown). We then determined the titers of replication-competent HIV-1_JRCSF preparations using JC.53 and TZM-bl cells as well as other representative HeLa-CD4/CCR5 clones in the JC panel. The results are plotted in Fig. ​Fig.1B1B as a function of cellular CCR5 content. Clones having more than a low threshold level of ∼8,000 CCR5/cell were equally susceptible to infection regardless of whether they contained MLV antigens, clearly demonstrating that HIV-1_JRCSF titers were not significantly affected by MLV. As expected, titers obtained with JC.53 and TZM-bl cells were also equivalent. In addition, these results demonstrate that HIV-1_JRCSF preparations made in JC.53 cells and in cells lacking MLV antigens (i.e., HEK293T cells and human peripheral blood mononuclear cells) were unable to infect HeLa cells lacking CCR5. The results in Fig. ​Fig.1B1B were expected because previous studies demonstrated that ecotropic MLVs cannot infect human cells or even bind to the human CAT-1 receptor paralog (1, 6, 21, 23). Moreover, it has been shown that ecotropic host range MLVs do not interfere with superinfection by any retrovirus capable of infecting human cells, including gibbon ape leukemia virus, amphotropic MLV, baboon endogenous virus, and feline leukemia virus subgroup C (21). In view of the report by Takeuchi et al. (22), we were surprised to find that JC.53 and TZM-bl cells express very small amounts of ecotropic MLV Env glycoproteins, as indicated by immunofluorescence microscopy and by their resistance to complement-dependent killing by a cytotoxic antiserum specific for MLV envelope glycoproteins (6). Nevertheless, the cell clones that contained MLV Gag all released ecotropic host range virions that replicated in murine NIH 3T3 cells but not in human cells (data not shown).To determine whether MLV affects the measurement of neutralizing Abs in TZM-bl cells, parallel assays were performed in TZM-bl and JC.48 cells; these latter cells were determined to be MLV free by Western blot analysis (Fig. ​(Fig.1)1) and by an inability to transfer MLV infection to NIH 3T3 cells (data not shown). Because JC.48 cells express CCR5 at somewhat lower levels than JC.53 cells (∼2-fold lower; Fig. ​Fig.1B),1B), it may be expected that they would be less susceptible to HIV-1 infection than are TZM-bl cells. Differences in susceptibility to HIV-1 infection may require the use of adjusted virus doses to achieve equivalent assay performance when measuring neutralizing Abs. Indeed, levels of CD4 and CCR5 were approximately twofold lower on JC.48 cells than on TZM-bl cells, whereas levels of CXCR4 were approximately equal (Fig. ​(Fig.1C).1C). We therefore measured the susceptibility of both cell lines to infection by three molecularly cloned Env-pseudotyped viruses, each bearing an Env from a different CCR5-tropic HIV-1 subtype B virus (SF162.LS, Bal.26, and QH0692.42). Infection was quantified by Luc activity expressed as relative luminescence units (RLU). Because JC.48 cells do not contain a reporter gene, the Env-pseudotyped viruses were prepared by cotransfection with the NL4-3.Luc.R-E- reporter backbone plasmid (7). Identical Luc-containing, Env-pseudotyped virus stocks were used in both cell lines. As shown in Fig. ​Fig.2A,2A, the infectivity of each pseudotyped virus was somewhat diminished in JC.48 cells compared to the infectivity in TZM-bl cells. Nonetheless, the levels of infectivity in JC.48 cells remained acceptable for neutralization assays.Open in a separate windowFIG. 2.HIV-1 infectivity and neutralization in TZM-bl and JC.48.CD4.CCR5 cells. (A) TZM-bl and JC.48 cells were incubated with serial fourfold dilutions (11 dilutions total) of three HIV-1 Env-pseudotyped viruses in quadruplicate in 96-well culture plates. Luc activity was measured after 48 h of incubation and is expressed as RLU after subtraction of background luminescence from cell control wells. Squares, TZM-bl cells; triangles, JC.48 cells. (B) Neutralization assays were performed with three HIV-1 Env-pseudotyped viruses in either TZM-bl or JC.48 cells. Input virus doses were adjusted to yield equivalent infectivity in both cell lines. Black bars, TZM-bl; gray bars, JC.48. Top panel: sCD4, monoclonal Abs, and HIVIG (purified immunoglobulin G from pooled HIV-1-positive plasmas). Bottom panel: individual HIV-1-positive plasma samples. The same three stocks of virus were used in both experiments. All three Env-pseudotyped viruses were prepared with the NL4-3.Luc.R-E- reporter backbone plasmid.With this information in hand, neutralization assays were performed in JC.48 and TZM-bl cells using adjusted virus doses that yielded equivalent infectivity levels in both cell lines. These neutralization assays were performed in a 96-well format as described previously (12), where the 50% inhibitory dose (ID₅₀) was reported as either the concentration or sample dilution at which RLU were reduced by 50% compared to RLU in virus control wells (cells plus virus without test sample) after subtraction of background RLU from cell control wells (cells only). A wide variety of serologic reagents was tested, including sCD4, a monoclonal Ab to the CD4 binding site of gp120 (immunoglobulin G1b12) (15); a monoclonal Ab that recognizes a glycan-specific epitope on gp120 (2G12) (5); two monoclonal Abs that recognize adjacent epitopes in the membrane proximal external region of gp41 (2F5 and 4E10) (2, 4); and serum samples from seven antiretroviral-naive HIV-1-infected individuals. As shown in Fig. ​Fig.2B,2B, results in the two cell lines were similar for all three viruses and all serologic reagents tested. Indeed, ID₅₀ values in the two cell types agreed within twofold, which is within the normal range of variability of the assay. These results indicate that equivalent neutralization results were obtained in both cell lines.In summary, we found no evidence that ecotropic MLV contamination in TZM-bl cells has a measurable effect on HIV-1 neutralization when these cells are used as targets for infection. This outcome indicates that the presence of ecotropic MLV in TZM-bl cells does not alter the ability of Ab to neutralize HIV-1, nor does it interfere with the detection of neutralization by using HIV-1 Tat-regulated reporter gene expression in a single-cycle infection assay. However, we discourage the use of TZM-bl cells to generate HIV-1 stocks, because the latter would likely be contaminated with ecotropic MLV and contain pseudovirions with mixtures of HIV-1 and ecotropic MLV Env glycoproteins. For this reason, we have begun efforts to produce an uncontaminated, second-generation panel of HeLa-CD4/CCR5 cell clones that express diverse amounts of CCR5 and to isolate a TZM-bl variant lacking MLV antigens.