To the Editor:
In a recent article in Blood, Yang et al1 reported that they found no evidence for abnormally elevated expression levels of CD95 ligand (L) mRNA in either fresh or cultured peripheral blood lymphocytes from 9 human immunodeficiency virus (HIV)-infected individuals with absolute T-cell counts of 200 to 500 cells/μL (CDC class 2) when compared with 6 uninfected healthy controls. These findings are in striking contrast to data previously published by us2 showing elevated CD95L mRNA levels in freshly isolated peripheral blood T cells from symptomatic HIV-1–infected children.
Yang et al1 used a sensitive quantitative reverse transcription-polymerase chain reaction similar to the one developed in our laboratory.2,3 However, these investigators used unseparated peripheral blood lymphocytes as the source for total RNA instead of isolated T cells. T, B, and NK lymphocytes differ with respect to spontaneous or stimulation-induced production of CD95L mRNA.3 In our hands, the striking difference in CD95L mRNA levels between HIV-infected children and noninfected controls could repeatedly be shown only in freshly isolated T cells.
Theoretically, the T-cell isolation procedure using immunomagnetic beads lasting up to 2 hours would allow for ongoing specific RNA synthesis to proceed. Therefore, blood samples from both patients and controls received exactly the same treatment. Thus, our data might only prove that CD95L mRNA synthesis is more easily inducible in T cells from HIV-1–infected children than in T cells from healthy controls. However, CD95L expression is increased in HIV-infected individuals also on the protein level. We performed sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting using a monoclonal mouse IgG antibody raised against a peptide corresponding to the extracellular part of the human CD95L molecule (Dianova, Hamburg, Germany) on cell lysates from purified T cells of a healthy control either freshly isolated or after 4 or 20 hours of stimulation with phorbol myristate acetate and ionomycin. The amount of CD95L protein detected in the lysate did not differ between unstimualted control cells and cells stimulated for 4 hours and increased significantly after 20 hours of stimulation. The lysate of unstimulated freshly isolated T cells from an HIV-1–infected patient contained at least as much CD95L protein as the control cells stimulated for 24 hours.
When following CD95L-mRNA levels in isolated T cells from 11 of our patients during 12 different courses of new antiretroviral treatment with reverse transcriptase inhibitors (zidovudine, didanosine, zalcitabine, and lamivudine, either alone or in combination), a significant reduction by 61% ± 8% (mean ± SEM; 1,441 ± 159 fg/250 ng RNA before v 564 ± 141 fg/250 ng RNA after 8 weeks of therapy; P < .001) was observed (Fig 1). At the same time point, an overall decrease of HIV viral load of 81% ± 12% was seen, whereas no consistent changes in the CD4+ T-cell counts could be noted during the short observation period after the start of therapy (mean change from baseline, −9% ± 8%). The reduction in CD95L-mRNA levels was more pronounced in patients exhibiting a good response to treatment (>1.0 log10 copies/mL reduction in HIV plasma viral load) than in poor responders (<1.0 log10 copies/mL reduction in HIV plasma viral load and/or progressive immunologic deterioration during the following 6 months). One patient showed a pronounced decrease in CD95L-mRNA levels and a decrease in viral load of approximately 1.0 log10 copies/mL. In this patient, serial determinations of CD95L-mRNA levels in T cells and HIV viral load in plasma showed a close relationship between both parameters (Fig 2). The initial decrease of viral load and CD95L-mRNA after starting therapy with zidovudine and didanosine was followed by an increase in viral load after 12 weeks that was paralleled by an increase in CD95L-mRNA levels and a concomitant decrease in CD4+ T-cell counts.
![Fig. 1. CD95L-mRNA levels in freshly isolated T cells from 11 HIV-1+ children before and after 8 weeks of therapy with new or additional reverse transcriptase inhibitors (12 treatment courses). Response to treatment was classified either according to the reduction in HIV-1 plasma viral load achieved at that time point or based on clinical and immunologic observations ([•] good response, <1.0 log10 copies/mL reduction in viral load; [○] poor response, <1.0 log10 copies/mL reduction in viral load). Mean CD95L-mRNA levels after 8 weeks of therapy were significantly lower than before therapy (P < .001; two-tailed paired Student's t-test). HIV-1 plasma viremia was determined using either a commercially available polymerase chain reaction kit (Amplicor Monitor; Roche Diagnostics, Grenzach-Whylen, Germany) or a sensitive second generation branched-chain DNA (bDNA) assay (Quantiplex HIV RNA 2.0 Assay; Chiron Diagnostics, Ferning, Germany) as recommended by the manufacturer.](https://ash.silverchair-cdn.com/ash/content_public/journal/blood/90/2/10.1182_blood.v90.2.886/4/m_bl_0059f1.jpeg?Expires=1735532942&Signature=vqxd8o-27c5kLo5hDmbR-2wNroa8YX1EUbp2k0WnJxHcy84hbPX5pgSbkwkqT8xSYMnA6HG~A2EoXD2SLnJP-ZTECrMJo~VATqsgAk01Wco8T3ygS72p4mbLwP1~yvm-whQ~BgGIUtrTMQp~pQdWOIB9-hGSnOKIpB5UWV09~Vvzkr82U2gvF6drrMxyDROGkANhLJ-mLzWsQ9W~onsQKd5AUfJ4MZCN0G3tUtMYyK3T0BoX2BwLKS-EWZjYr~yieIx0QKQEldy5tN2KjBNAySWwyOV9SylFazX2Sy1WV1WxD2BCKRoI0lS~bK1jYER56oMFUNJYtCy~6y8kOfoGIA__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
CD95L-mRNA levels in freshly isolated T cells from 11 HIV-1+ children before and after 8 weeks of therapy with new or additional reverse transcriptase inhibitors (12 treatment courses). Response to treatment was classified either according to the reduction in HIV-1 plasma viral load achieved at that time point or based on clinical and immunologic observations ([•] good response, <1.0 log10 copies/mL reduction in viral load; [○] poor response, <1.0 log10 copies/mL reduction in viral load). Mean CD95L-mRNA levels after 8 weeks of therapy were significantly lower than before therapy (P < .001; two-tailed paired Student's t-test). HIV-1 plasma viremia was determined using either a commercially available polymerase chain reaction kit (Amplicor Monitor; Roche Diagnostics, Grenzach-Whylen, Germany) or a sensitive second generation branched-chain DNA (bDNA) assay (Quantiplex HIV RNA 2.0 Assay; Chiron Diagnostics, Ferning, Germany) as recommended by the manufacturer.
CD95L-mRNA levels in freshly isolated T cells from 11 HIV-1+ children before and after 8 weeks of therapy with new or additional reverse transcriptase inhibitors (12 treatment courses). Response to treatment was classified either according to the reduction in HIV-1 plasma viral load achieved at that time point or based on clinical and immunologic observations ([•] good response, <1.0 log10 copies/mL reduction in viral load; [○] poor response, <1.0 log10 copies/mL reduction in viral load). Mean CD95L-mRNA levels after 8 weeks of therapy were significantly lower than before therapy (P < .001; two-tailed paired Student's t-test). HIV-1 plasma viremia was determined using either a commercially available polymerase chain reaction kit (Amplicor Monitor; Roche Diagnostics, Grenzach-Whylen, Germany) or a sensitive second generation branched-chain DNA (bDNA) assay (Quantiplex HIV RNA 2.0 Assay; Chiron Diagnostics, Ferning, Germany) as recommended by the manufacturer.
![Fig. 2. HIV-1 viral load in plasma ([•] 104 copies/mL), absolute CD4+ T-cell counts ([▵] cells per microliter), and CD95L-mRNA levels in freshly isolated T cells ([○] femtograms per 250 ng total RNA) during antiretroviral combination therapy with zidovudine and didanosine in an HIV-1+ infant. Initial reduction of HIV-1 viral load and CD95L-mRNA levels were paralleled by an increase in CD4+ T-cell counts. However, as early as 12 weeks after the introduction of antiretroviral treatment, HIV-1 viral load and CD95L-mRNA levels increased and CD4+ T-cell counts decreased.](https://ash.silverchair-cdn.com/ash/content_public/journal/blood/90/2/10.1182_blood.v90.2.886/4/m_bl_0059f2.jpeg?Expires=1735532942&Signature=gAT78RDl-upEHqTlNZppXXNuR1c5qmWDn8TuT~bAVsNnF086OaepXvL9kxjBhDpUyOLb2FTKce~tk0osZz4pJKYgfq9rt0rp847LBOZptlldfDlhRyhf7RVKWm3a2vEz3YE4eQAqyQeheeHxY-SzrODVg7PyGRJnzVlR9fHiNRFzRtGQxULH2HdAVEPjoyeiGgF5QKf2ezOSVWi91g3Y3N8~Zc1uyNd8KCZpKr7-zKp7lAlbc-HMksJA7tq2ZgbAK4PCXNyBAw1gEASuznkpbdt-4bFEDIXpNYmp8xc0sMoLv8Kn6BfG1BNdbwcGf45SpvfVY5iVKa0TGD5bTnz7GQ__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
HIV-1 viral load in plasma ([•] 104 copies/mL), absolute CD4+ T-cell counts ([▵] cells per microliter), and CD95L-mRNA levels in freshly isolated T cells ([○] femtograms per 250 ng total RNA) during antiretroviral combination therapy with zidovudine and didanosine in an HIV-1+ infant. Initial reduction of HIV-1 viral load and CD95L-mRNA levels were paralleled by an increase in CD4+ T-cell counts. However, as early as 12 weeks after the introduction of antiretroviral treatment, HIV-1 viral load and CD95L-mRNA levels increased and CD4+ T-cell counts decreased.
HIV-1 viral load in plasma ([•] 104 copies/mL), absolute CD4+ T-cell counts ([▵] cells per microliter), and CD95L-mRNA levels in freshly isolated T cells ([○] femtograms per 250 ng total RNA) during antiretroviral combination therapy with zidovudine and didanosine in an HIV-1+ infant. Initial reduction of HIV-1 viral load and CD95L-mRNA levels were paralleled by an increase in CD4+ T-cell counts. However, as early as 12 weeks after the introduction of antiretroviral treatment, HIV-1 viral load and CD95L-mRNA levels increased and CD4+ T-cell counts decreased.
We conclude from our data that constitutive expression of CD95L is increased in freshly isolated peripheral blood T cells from HIV-1–infected children. Studies addressing CD95L mRNA expression in HIV-infected individuals should use freshly isolated lymphocyte subpopulations rather than bulk peripheral blood lymphocytes as the RNA source. Cryopreservation of mononuclear cells or any delay in processing the blood and RNA samples should be avoided because these procedures usually result in a decrease rather than an increase in measurable CD95L mRNA levels (I.H. and K.-M.D., unpublished observations).
The data presented here also indicate that increased CD95L production in freshly isolated T cells from HIV-1–infected children can be downregulated to near-normal levels by effective antiretroviral treatment. The CD95 receptor/ligand system critically regulates physiological elimination of T cells.4 In peripheral T cells, activation-induced expression of CD95L mediates apoptosis after binding to CD95L on the surface of the T cell. Upregulation of CD95L in (uninfected) peripheral blood T cells from HIV-1+ individuals may be either caused specifically by cellular uptake of HIV-1 Tat5 or may reflect general stimulation of lymphomonocytic cells observed during HIV infection.6,7 In addition to T cells, infection of human macrophages with laboratory HIV strains in vitro leads to an upregulation of CD95L expression in these cells that caused apoptotic death of uninfected T lymphocytes.8 We previously showed inhibition of CD95-mediated apoptosis in T cells from HIV-infected children by disrupting CD95 receptor/ligand interaction using blocking anti-CD95 F(ab′) antibody fragments.2 Given the important role of CD95L in T-cell apoptosis, we speculate that effective antiretroviral treatment (including HIV protease inhibitors) that leads to significant reduction of virus replication will also lead to decreased activation-induced death of uninfected T cells, thus restoring impaired T-cell homeostasis and function. Furthermore, monitoring CD95L mRNA expression in T cells may give important information on self-destructive immune stimulation in HIV-infected individuals and might be used as an additional surrogate marker for disease progression and treatment response.
Response
To the Editor:
Böhler et al have pointed out a discrepancy between our respective studies1-1,1-2 with regard to spontaneous expression of Fas ligand (FasL; or CD95L) by peripheral blood cells from HIV+ individuals. We did not address their findings in our report, which was in press at the time their report was published, but are happy to have the opportunity to do so now. We used quantitative competitive polymerase chain reaction to analyze FasL levels in unfractionated peripheral blood mononuclear cells (PBMC) from HIV-infected adult patients with absolute CD4 counts of 200 to 500/μL. We purposely chose to analyze unfractionated PBMC because non-T cells, such as NK and monocytes, can express biologically active FasL and therefore participate in the fratricide of Fas+ T cells.1-3,1-4 We found that FasL mRNA levels in freshly isolated blood were comparable between HIV+ and HIV− individuals and increased to similar extents in both groups after overnight incubation at 37°C. Spontaneous ex vivo death of peripheral blood lymphocytes from HIV+ patients was completely inhibited by a soluble chimeric form of Fas, demonstrating a requirement for Fas/FasL interactions in the death of cultured cells. Using a similar technique to study only purified T cells, Bäumler et al1-1 found that FasL levels were spontaneously elevated in a subset of HIV+ pediatric patients. Curiously, despite this finding, in their hands a blocking anti-Fas antibody had no effect on spontaneous apoptosis in medium and only partially blocked activation-induced apoptosis. Nevertheless, the investigators reached the same conclusion as we, ie, that upregulated FasL and Fas contribute to the death of T cells from HIV+ individuals.
Böhler et al argue that our analyses did not detect spontaneously elevated levels of FasL because we examined unseparated PBMC rather than T cells. We do not believe this to be the case. Antihuman FasL monoclonal antibodies are now available, making it possible to follow FasL protein expression on cell subpopulations in the peripheral blood. Two of us have used three-color flow cytometric analysis to assess FasL levels on fresh PBMC from antiviral drug-treated adult HIV+ patients with absolute CD4 levels of less than 500/μL (H.N. and S.P., manuscript in preparation). These studies have shown that the levels of FasL on fresh lymphocytes from HIV− donors (n = 9) are very low (Δ mean fluorescence intensity of 2.1 ± 1.5 as compared with isotype controls). Of note is the finding that the FasL levels on lymphocytes from HIV+ donors (n = 9) were as low or even lower (Δ mean fluorescence intensity of 0.8 ± 0.9). It is also noteworthy that the constitutive level of FasL on monocytes from both HIV+ and HIV− PBMC was approximately threefold higher than that on lymphocytes. Thus, monocytes cells may represent a biologically significant source of FasL in PBMC. These independent results confirm the finding that FasL levels are not spontaneously elevated on T cells of adult HIV+ patients.
How, then, can one reconcile the differences between the studies with regard to FasL expression? There are at least two major factors that may account for this discrepancy. First and foremost, there was a substantial difference in the medical management of the patients. All of our patients were on two antiviral drugs, AZT and didanosine. In contrast, most of the children in the study by Bäumler et al1-1 received only monthly Ig infusion, and only 9 of 24 children received AZT single-drug therapy. As Böhler et al have pointed out in their letter, the spontaneously elevated levels of FasL in the T cells from these children decreased when more vigorous retroviral therapy was initiated. Therefore, aggressive retroviral therapy of the patients in our study may have suppressed T cell FasL expression, in which case there is no real disagreement between the studies. Second, the patient populations were quite different: we analyzed cells from adult patients, whereas Bäumler et al1-1 studied samples from children. Taken together, it seems to us that the two studies agree on the major issue: ex vivo death of peripheral blood lymphocytes from HIV+ individuals is due to FasL binding to constitutively expressed and biologically active Fas. Regardless of whether FasL is elevated constitutively or as a consequence of activation by specific antigens or mitogens, it is an intriguing target for possible therapeutic intervention. As we have shown, retinoic acid1-5 and the calpain inhibitor E64d2 inhibit activation-induced upregulation of FasL, and both inhibit the ex vivo apoptosis of PBMC from HIV+ donors.1-2 1-6 Clinical trials with these or other reagents that interfere with Fas-mediated signaling will be necessary to determine if blockade of this death effector pathway has a beneficial effect on the progression of T-cell dysfunction and loss in acquired immunodeficiency syndrome.
