Rathbun et al1 have reported apoptosis in the Fanconi anemia (FA) cell line HSC536N in response to interferon-γ (IFN-γ) and agonistic Fas antibodies. While they did not use the classic assays of apoptosis (morphology, TUNEL, DNA laddering), they concluded that cell death was due to apoptosis because of evidence of both caspase 3 and PARP cleavage on Western blots. In order to show cleavage, however, they had to load 100 μg of protein per lane onto the gel and, allowing for overloading, the blots appear to show substantially less than 1% cleavage. In our experience, substantial cleavage of caspase 3 during apoptosis can be seen at standard loadings of 10 μg or less, and PARP cleavage is often quantitative in nature. Since we and others have shown that FA cell lines do not undergo normal apoptosis in response to mitomycin C,2 3 we were interested whether the reported conclusions were correct.
We incubated the FANCC mutant cell line HSC536N for 6 hours with different doses of IFN-γ, with or without an agonistic Fas antibody, and measured cell death by 7-amino actinomycin D (7AAD) staining and flow cytometry, which detects early changes in membrane permeability.4 Surprisingly, we found no response to IFN-γ alone and no enhancement of Fas-mediated killing (Table1).
Cell extracts were subjected to Western blotting and probed with an antibody for caspase 3 (Pharmingen, catalog no. 65906E, San Diego, CA) (Figure 1). In contrast to Rathbun et al, we found substantial cleavage of caspase 3, with generation of p20 and p17 cleaved forms, with only 10 μg protein loading. To further investigate the discrepancy with the previous report, we repeated the blots using the same antibody employed by Rathbun et al (Transduction Laboratories, C31720, San Diego, CA). Cleaved forms of caspase 3 were scarcely detectable at 10 μg loading (Figure 1B). We note that the manufacturer only claims detection of the proform of caspase 3 by this antibody. Western blots were also performed with an anti-PARP antibody. In this case we found no detectable evidence of PARP cleavage, compared to standard controls (Figure 1). DNA was also extracted for gel electrophoresis, and DNA laddering was not detected.
Western blotting to detect cleaved forms of caspase 3 and PARP.
(A) HSC536N cells (C) and Jurkat cells (J) were treated with anti-Fas antibody for 6 hours and analyzed for the p20 and p17 cleaved forms of caspase 3 (top), and the p85 cleaved form of PARP (bottom). Note an artefact band migrating faster than p85 PARP is present in treated and untreated cells. (B) Jurkat cells were treated with etoposide for 6 hours and analyzed for the p20 and p17 cleaved forms of caspase 3 using Pharmingen antibody catalog no. 65906E1 or Transduction Laboratories antibody C31720.2
Western blotting to detect cleaved forms of caspase 3 and PARP.
(A) HSC536N cells (C) and Jurkat cells (J) were treated with anti-Fas antibody for 6 hours and analyzed for the p20 and p17 cleaved forms of caspase 3 (top), and the p85 cleaved form of PARP (bottom). Note an artefact band migrating faster than p85 PARP is present in treated and untreated cells. (B) Jurkat cells were treated with etoposide for 6 hours and analyzed for the p20 and p17 cleaved forms of caspase 3 using Pharmingen antibody catalog no. 65906E1 or Transduction Laboratories antibody C31720.2
Are these cells really undergoing apoptosis? The primary definition of apoptosis is based on morphology. Examination of Giemsa-stained cytocentrifuge preparations of anti-Fas–treated cells clearly showed classical apoptosis, with nuclear condensation, and chromatin marginalization and fragmentation. We conclude that HSC536 cells do indeed undergo morphological apoptosis in response to Fas antibodies and that the failure of Rathbun et al to detect quantitative caspase 3 cleavage was a technical problem associated with the antibody. On the other hand, even if the minor band seen in their blots is a PARP cleavage product, cleavage of PARP within these cells is still negligible.
There is clear evidence that primary FA cells are hypersensitive to IFN-γ and that HSC536 cells show biochemical abnormalities in their response to IFN-γ. However, we could not detect any enhancement of Fas-mediated cell death by IFN-γ.5 We suggest this is because of the high constitutive level of Fas expression on these cells.
Apoptosis and FANCC
Rutherford et al have stated that in our paper on caspase activation in Fanconi anemia (FA)1-1 cells “the classic assays of apoptosis” were not used. They are incorrect. Figure 3 was derived from experiments using the TUNEL assay. An article by our group that was published in Blood a year earlier1-2also reported our experience using both morphology and TUNEL assays. In both papers we demonstrated that FA cells are excessively apoptotic. Since the publication of the paper in question we have published additional papers that have further clarified the molecular nature of the apoptotic FA-C phenotype1-3-1-8 and have identified a pivotal control point for caspase 3 activation in FA cells.1-6 1-7
IFN itself cannot fully reveal a differentially hypersensitive FA phenotype in the isogenic cell lines. We reported this a number of years ago when we noted that a blocking antibody to fas abrogated apoptotic responses of FA-C cells to IFNγ indicating that fasL in cocultured cells contributed to what appeared to be simply an IFN effect.1-2 Later we reported that the combinations IFNγ and fasL, IFNγ and TNFα, or IFNγ and dsRNA all reveal exaggerated responses in FA-C cells1-6-1-8; that the double-stranded RNA-dependent protein kinase, PKR, is constitutively activated and excessively induced by cytokines in FA cells1-6; and that the role played by FANCC in modulating PKR activity depends upon its capacity to bind to hsp70, a known modulator of PKR activity.1-7 In fact, PKR is a bottleneck point for the apoptotic pathways because a dominant negative inhibitor of PKR abrogates all differential cytokine responses between FA cells and normal cells.1-6 Because IFNγ induces expression of PKR, and TNF/fas/dsRNA activate the molecule, IFN functions as an enhancer of PKR activity.
In our hands, the most reliable and quantitative assessments of the induced apoptotic phenotype in FA cells (murine or human cells) are made using a flow cytometric method that permits quantification of single cells containing caspase 3.1-6-1-8 Using this method, TNFα, fas-ligand, and double-stranded RNA induce more apoptotic responses in FA cells pretreated with IFN than in normal cells treated with IFN.1-1 1-6-1-8 In view of the consideration by Rutherford et al that “classic assays” for apoptosis include TUNEL and morphological analyses, we are not sure why they chose to use 7-AAD staining as their litmus test.
Rutherford et al have stated that “in order to show cleavage,” our group “had to load 100 μg of protein.” This, too, is incorrect. At no time have we stated that detection of cleavage of either PARP or caspase 3 required a 100 μg load. Also, the authors' statements that “in contrast to Rathbun et al, we found substantial cleavage of caspase 3…with only 10 μg loading,” and “failure of Rathbun et al to detect quantitative caspase 3 cleavage…” are misleading. Their observation does not stand in contrast to ours. We consistently find differential caspase 3 activation and over the past several years have shown that caspase 3 activation can be shown at a single-cell level, a quantitative standard that simply cannot be matched by Western blotting. Indeed, the result reported here by Rutherford et al at once confirms our results and clarifies that the anti-PARP antibody in question is capable of detecting the cleaved form of the molecule. While we have not seen differential PARP activation in isogenic cell lines, we view the immunoblot method as insensitive and have not focused much attention on PARP recently because we now recognize that PARP cleavage is not necessary for execution of an apoptotic program.1-9
The authors' claim that FA cell lines treated with MMC do not undergo “normal apoptosis” is difficult to understand in part because the only cited reference that used isogenic FA-C cells did, in fact, demonstrate clearly that PARP cleavage induced by MMC was noted only in mutant cells, not in FANCC corrected cells.1-10 (Fig5C) Moreover, work from our laboratory published in this journal shows a major reduction in TUNEL-positive MMC-exposed cells in FANCC-complemented FA-C cells.1-2 (Fig6A) Of equal relevance we have also noted (G.C.B. and R. K. Rathbun, unpublished observations, September 2000) that MMC-exposed FA cells have more cleaved caspase 9 than do MMC-exposed FA-C cells complemented with FANCC.
Because there is not much disagreement on the involvement of caspase 3 in the FANCC pathway, it is probable that all of us share the view that control points for caspase 3 activation are critically important for an understanding of FA cells. We have chosen this path.1-1,1-3 1-5-1-8 What are we left with? We are possibly left with some evidence that internucleosomal cleavage of chromatin is somehow different in FA cells treated with MMC, but we are not left with any viable notion that MMC does not induce a fully loaded apoptotic response.
In summary, Rutherford et al have attributed to our report specific deficiencies that did not exist. If their initial intention was to reproduce our work, their group should have done the same experiments. In fact, they failed to reproduce conditions we used (TUNEL positivity is most substantial after 48 hours of exposure to IFN and agonistic fas antibody1-2 [Fig3]), failed to use isogenic sets of FA mutant cells, and used a method for quantifying apoptotic cells that is not fully validated for studies on Fanconi cells, when compared to results from a battery of conventional measures of apoptosis, methods they themselves claim to be gold standards. Their concluding suggestion, “high constitutive levels of fas expression on FA cells” may account for some effects we have all seen, has been tested by us years ago and ruled out at least in B-cell lines.1-4 When the authors test their next hypothesis, they should utilize isogenic FA cell lines now widely extant, use gold standard apoptosis assays, use the combinations of cytokines now known to induce maximal differential apoptosis in FA cells, and consider using caspase 3–activation assays that can detect biologically meaningful fractional changes in cell populations. If they do these things, we have no doubt they will find that FA-C cells exhibit exaggerated responses that are both statistically and biologically significant. We suppose that one might spend a good bit of time arguing about certain semantic issues that have evolved in this field, issues that might be likened to the tree-falling-in-the-forest-unseen argument (eg, if all other molecular events characteristic of apoptosis occur, does a reduced detectability of internucleosomally cleaved DNA mean that there is no “apoptosis?”). Moving beyond such arguments, we are confident that most workers in this field, including the 2 parties disagreeing here, will ultimately concur that FA-C cells are more poised to undergo programmed cell death when challenged with certain chemical agents or selected extracellular biologic cues and that the apoptotic response of these cells is generally exaggerated because the protein encoded by the FANCC gene modulates such responses.
References
Supported by a grant from the United Kingdom Leukaemia Research Fund
