Possible pitfalls investigating cell death responses in genetically engineered mouse models and derived cell lines.

Genetically engineered mouse models are frequently used to identify pathophysiological consequences of deregulated cell death. Targeting pro-apoptotic or anti-apoptotic proteins of the extrinsic or intrinsic apoptotic signalling cascade is state of the art since more than two decades. Such animal models have been increasingly made use of over the past years to study loss- or gain-of-function consequences of one or more components of the molecular machinery leading to cell death. These studies have helped to separate redundant from non-redundant functions of apoptosis-related proteins in normal physiology and sometimes unravelled unexpected phenotypes. However, correct interpretation of data derived from knockout mice or derived cells and cell lines is often flawed by the comparison of cells originating from different inbred or mixed genetic backgrounds. Here we want to highlight some basic problems associated with genetic background-based modulation of cell death sensitivity and describe some methods that we use to investigate cell death responses in hematopoietic and non-hematopoietic cells. Thereby, we show that hematopoietic cells derived from wild type mice on a C57BL/6:129/SvJ recombinant mixed genetic background are significantly more resistant to spontaneous cell death or DNA-damage induced apoptosis in vitro than cells derived from inbred C57BL/6 mice. Furthermore, we show as an example that C57BL/6 mice are more susceptible to γ-irradiation induced cell death after whole body irradiation in vivo and subsequent T cell lymphomagenesis.

Introduction

Programmed cell death is critical for the correct development of all metazoans. The most prominent mode of programmed cell death in vertebrate cells is referred to as the “intrinsic” or “mitochondrial” pathway of apoptosis, initiated by loss of mitochondrial integrity, a process also known as mitochondrial outer membrane permeabilization (MOMP). This process is controlled by the complex interplay of different members of the B cell lymphoma-2 (BCL-2) protein family, that all orchestrate activation of the largely redundant pro-apoptotic effector proteins BAX and BAK. As a consequence of MOMP, members of a second protein family, essential for apoptosis, the cysteine-driven aspartate-specific proteases (caspases), i.e. caspase-9 and caspase-3, -6, -7 are activated and lead to the processing of downstream targets, ending in and mediating morphological features of apoptotic cell death (see for review [1]). The second apoptotic cell death pathway, the “extrinsic” pathway or death receptor (DR) pathway involves ligand-dependent activation of membrane receptors of the tumor necrosis factor receptor 1 (TNF-R1) superfamily, which in turn triggers the formation of the death inducing signalling complex (DISC), activating caspase-8 and in humans also caspase-10. Both pathways meet at the level of caspase-3 activation [2]. Not surprisingly, the activation of caspases is carefully controlled to avoid unwanted induction of apoptosis and deregulation of programmed cell death contributes to the pathogenesis of several human diseases including autoimmunity and cancer [3]. Accordingly, genetically engineered mouse models have been used increasingly to investigate the molecular mechanisms of apoptotic cell death and to study consequences of defective apoptosis in different disease settings, e.g., to gain new insights into carcinogenesis and the effectiveness of antitumor therapies and the molecular basis of drug-resistance.

Targeting proteins critical for cell death, including caspases, death receptors or BCL-2 family proteins in genetically engineered mouse model has been commonly used to study the pathophysiological consequences of impaired apoptosis at the level of the whole organism, but often also in cell lines, most prominently MEF or primary cells such as lymphocytes, hepatocytes or neurons derived from such mouse mutants. So far, mouse models targeting all known murine caspases (caspase-1, -2, -3, -6, -7, -8 and -9, -11, -12; see for review [4-7] and mouse models targeting members of the BCL-2 family have been generated and are still intensively used today [8-11].

Impact of genetic background on cell death susceptibility

The number of mouse strains used to establish embryonic stem cell lines was initially limited and the most commonly used embryonic stem cell lines were derived from the 129/SvJ inbred strain [12]. Chimeras are usually produced by injection of manipulated ES cells into blastocystes derived from C57BL/6 mice that are finally implanted into pseudopregnant outbred foster mothers. Thus, the resulting pups have tissues build up from cells derived from different genetic origins, i.e. from the embryonic stem cells of 129/SvJ mice and the inner cell mass of the C57BL/6 donor blastocyst. Male chimaeras are subsequently crossed with C57BL/6 (B6) females to screen for germ line transmission of the mutated allele in Agouti coloured offspring (F1). Usually heterozygous mutant mice should be backcrossed onto one of the two possible genetic backgrounds (usually C57BL/6) for at least eight to ten generations to avoid comparison of mice with different contribution of either background in F1 offspring that give rise to a recombinant inbred strain in F2 and all their possible descendants (F3, F4…). However this is not always performed vigorously enough to save time. In such recombinant inbred strains, the contribution of each individual background can vary significantly between siblings and litters, affecting all subsequent analysis. A number of reports have been published in the past demonstrating the strong influence of genetic background onto the results from transgenic or knockout mouse models. For example, caspase-3 deficient mice were described to die at an age of 1–5 weeks, were born at lower Mendelian frequency and of smaller size on a mixed C57BL/6:129/SvJ background [13,14] while Casp3 mice on a pure C57BL/6 background reach adulthood, are fertile and do not show the dramatic exencephalic phenotype noted on the mixed genetic background ([15] and own unpublished observations).

Loss of the tumor suppressor p53 on a 129/SvJ:Ola background manifests in a shorter latency period to tumor development and allows development of a small number of pituitary adenocarcinomas not seen on other genetic backgrounds [16]. Furthermore, on a mixed genetic background (129/SvJ × C57BL/6) approximately 9% of female p53−/− embryos show a defect in neuronal tube closure, compared to >90% penetrance of this phenotype on a pure C57BL/6 background [17-19]. In contrast, p53-deficency on inbred C57BL/6 background shows severe ocular abnormalities, in contrast to mice on the 129/SvJ background [20]. Bcl-2-deficient mice show a severe phenotype manifested in polycystic kidney disease and lymphopenia on a C57BL/6 background causing their death within the first 6 weeks of life, while Bcl-2 on a 129/SvJ background have a much milder phenotype [21]. Furthermore, mice lacking the BH3-only protein Bad on 129/SvJ background were reported to develop diffuse large B cell lymphomas with an incidence of about 20% but no such phenomenon was observed after backcrossing onto C57BL/6 [22,23] and the severe autoimmune pathology noted in Bim-deficient mice on a 129SvJ:B6 background [9,24] or Fas-deficiency in the MRL background is clearly strongly ameliorated upon backcrossing onto C57BL/6 [25]. Hence, the choice of a uniform genetic background and appropriate controls, if not littermates, is most critical to allow direct comparison of findings made in different mouse mutant strains. However, it has to be noted that even extensive back-crossing does not allow a completely clean transfer of a targeted allele onto the desired genetic background as genomic regions in close proximity of the targeted allele will maintain their original genetic identity, as proximity will prevent meiotic crossing over in these regions. In addition, extensive “in-house” breeding bares the danger of introducing genetic drifts into individual strains of mice kept in isolation without recurrent “infusion” of the original genetic background. However, given the extra time needed to “refresh” the original genetic background of nowadays often multi-allelic mouse mutant strains, this problem is often ignored.

To illustrate some of these problems, we noted a drastic difference in the mean lifespan of wild-type (wt) mice of different genetic backgrounds but also within the same tumor-prone mouse strain over time in our own colonies. A cohort of male mice maintained on an uncontrolled recombinant mixed C57BL/6:129/SvJ (1:1) background died due to age-related diseases with a median of 494 days (n = 10). Necropsy suggested severe inflammation of the urogenital tract (n = 2), hepatocellular carcinoma and/or kidney tumors (n = 4), lymphoma (n = 1), gastrointestinal tumor (n = 1) or stroke (n = 2), as cause of death, while male mice on a pure C57BL/6 background showed a mean disease-free survival of 730 days (n = 7), the defined experimental cut-off (p = 0.001; Logrank Mantel–Cox analysis; Fig. 1A). Yet, in a model of split-dose γ-irradiation induced thymic lymphomagenesis mice of mixed genetic background showed significantly delayed onset of disease (Fig. 1B). Briefly, 4-week old C57BL/6 wt mice were irradiated with a low dose of 1.75Gy 4-times in weekly intervals. Wild-type mice start to develop thymic lymphomas with a median onset of 183 days (n = 10; 7 female and 3 male animals), while mice on a mixed genetic background (C57BL/6:129/SvJ) showed a significantly longer (p = 0.0017) median tumor free survival of 260 days (n = 11; 8 female and 3 male; Fig. 1B) that was most prominent in female mice (p = 0.0029) but not significant in males, with the caveat of very low numbers. In relation to genetic drift caused by in-house breeding, we noted a significantly delayed onset of B cell lymphomagenesis in a cohort of C57BL/6J Eμ-Myc transgenic mice monitored in the years 2008–2010, (mean latency 202 days) when compared to earlier cohorts published with a mean latency of 138 days [26]. Strikingly, re-infusion of wt C57BL/6 background into the Eμ-Myc transgenic line reduced tumor latency again to 135 days in a new cohort monitored in 2011–2012 (Fig. 1C). Notably this genetic drift appeared to affect mainly male Eμ-Myc mice (Fig. 1D).

Fig. 1

Kaplan–Meier analysis reveals reduced life-span of recombinant inbred wild type (C57BL/6:129/SvJ) mice and genetic drift in in-house bred Eμ-Myc mice. (A) Cohorts of C57BL/6-wt (n = 7) and mixed C57BL/6:129/SvJ-wt mice (n = 10) were monitored over time and killed when mice showed signs of illness. A cut-off was made after 730 days. (B) Cohorts of wild type mice from C57BL/6 (n = 13; median 180 days) and C57BL/6:129/SvJ (n = 11; median 230 days) mice were included into the protocol of fractionated irradiation and the onset of thymic lymphoma was monitored over time by daily inspection. (C) Cohorts of Eμ-Myc transgenic mice monitored between 2008 and 2010 (n = 76) and 2011 and 2012 (n = 41) revealed significant differences in tumor latency (mean 202 vs. 135 days). (D) Delayed disease onset was only observed in male mice in the 2008–2010 cohort (n = 42) vs. 2011–2012 cohort (n = 34). Females did not show significant differences (p = 0.149; n = 34 vs. n = 22; not shown). Tumor free survival is depicted as a Kaplan–Meier plot. Significances were defined by Logrank (Mantel–Cox) analysis.

Analysing the distribution of immune cell subsets in bone marrow, thymus or spleen from wt mice on C57BL/6:129/SvJ (1:1) mixed genetic background showed significantly reduced total cellularity in the spleen compared to inbred C57BL/6 mice (p = 0.047). The difference observed was mainly due to lower numbers of B220+ B cells, affecting all stages of development, including transitional type 1 (T1), type 2 (T2) and follicular (FO) B cells (Fig. 2A–D).

Fig. 2

Reduced number of splenic B cells in C57BL/6:129/SvJ mice. In order to analyse the distribution of B and T cell subsets, we have isolated single cell suspensions from thymus (A), bone marrow (BM) (B) and spleen (C and D) by either mashing the organ (spleen, thymus) through a 100 μm cell strainer (BD, Austria) or by resuspending bone marrow. After washing the cells in 1× PBS containing 10% FCS (PAA), Pen/Strep (Sigma–Aldrich) and Gentamycin (Gibco) for 5 min at 500g at 4 °C cell pellet was resuspended in 1× PBS including supplements mentioned before and total organ cellularity was assessed by counting cells in a chamber hemocytometer (Neubauer). To investigate the distribution of B and T cell subsets aliquots of cell suspensions were stained using fluorochrome-conjugated (FITC, PE, APC or biotin) cell surface marker-specific rat-anti-mouse antibodies: (1) T cells (thymus and spleen): CD4 (GK1.5) CD8 (YTS169) or pan-T cell marker Thy1/CD90 (T24.31.2), (2) B cells: (a) bone marrow: IgM (5.1), IgD (11/26C); CD-43 (R2/60), IgM (5.1) and B220 (RA3-6B2); (b) spleen: IgM (5.1), IgD (11/26C), CD21 (7G6, BD) and CD23 (B3B4, BD); Thy1 and B220 (RA3-6B2). All antibodies were purchased from eBioscience. After incubation of cells for 20 min at 4 °C cells were washed for 5 min at 500g at 4 °C and, if needed, cells were incubated for additional 20 min with a relevant secondary antibody (streptavidin-R-PE from DAKO or streptavidin-vPE-Cy7 from BD). After the final washing step cells were resuspended in 250 μl of 1× PBS and flow cytometric analysis was performed using a FACScalibur cell analyser (BD, Austria). Pro B cells: B220+CD43+; Pre B cells: B220+CD43−IgM−; T1 B-cells: IgMhigh CD21−; T2 B-cells: IgMhighCD21high; MZ (marginal zone B-cells): CD21+CD23low; FO (follicular B-cells): CD21+CD23+. Data are presented as means ± SEM of >3 animal per genotype.

So far we did not further specify which type of C57BL/6 mice were used, as there are several substrains available that are not genetically identical and for which behavioural differences have been reported [27]. The two most commonly used substrains are C57BL/6J and C57BL/6N. The original Jax strain is C57BL/6J and the 6N substrain was derived from it at the NIH and separated in 1951 [28]. The major difference between these lines is deficiency for the nicotinamid nucleotide transhydrogenase (Nnt) gene in the C57BL/6J substrain. Nnt is a pyridine nucleotide transhydrogenase and integral inner mitochondrial membrane protein. It is part of the energy-transfer system of the respiratory chain and catalyzes the transfer of a hydride ion between nicotinamide adenine dinucleotide, NAD(H), and oxidized nicotinamide dinucleotide phosphate, NADP(H). A direct comparison of cell death sensitivity between C57BL/6J and C57BL/6N strains has not been performed but as the enzyme contributes to antioxidant defense, differences in tolerance to ROS can be expected. Loss of Nnt was recently described as a cause for familial glucocorticoid deficiency in humans and increased levels of active caspase-3 were found in the adrenal glands from C57BL/6J mice when compared to the C57BL/6N strain [29]. Knockdown of NNT in a human adrenal cell line also caused an increase in ROS levels [29]. These findings certainly pose a caveat to the direct comparison of cell death mutant mouse strains or derived cells from mice maintained on different or mixed C57BL/6 backgrounds and prompted us to define the C57BL/6J and C57BL/6N status in our colonies with the result that we occasionally compared mutations on C57BL/6J with others on mixed C57BL6J/N backgrounds in our studies. Both strains can be easily discriminated by PCR genotyping as a genomic region of >18 kb, containing exons 7–11 of the Nnt gene locus, is missing in the C57BL/6J strain [30]. Primers covering the deletion are positioned in exon 6 (5′GTAGGGCCAACTGTTTCTGC3′) and exon 12 (5′TCCCCTCCCTTCCATTTAGT3′) and identify a C57BL/6J contribution by giving rise to a 547 bp PCR product. To identify a C57BL/6N contribution, we use published exon 8 specific primers (fwd: 5′CCAGGCGAGCACTCTCTATT3′ and rev: 5′CAGGGTCACAGGAGAACACA3′) giving rise to a 182 bp PCR fragment using the following conditions (94 °C for 2 min, then 30 cycles 94 °C 15 s, 58 °C 20 s, 72 °C 30 s) [31]. In addition to the Nnt deletion an increasing number of SNPs (currently about 30) has been identified in the different C57BL/6 lines that may affect gene expression. For details see the Mouse Genome Database (MPD): phenome.jax.org/SNP/.

Together, these observations underline the importance of controlled genetic background choice prior analysis and document again its strong influence on different spontaneous as well as induced pathologies that associate with deregulation of apoptotic cell death.

Isolating individual lymphocyte subsets to study apoptosis ex vivo

Keeping above-mentioned findings in mind one can easily envision that such differences extrapolate also to experiments in tissue culture where cell death phenotypes are explored in primary or immortalized cells derived from mouse mutants lacking or overexpressing different apoptosis regulators. To investigate the role of pro-apoptotic proteins on programmed cell death in the hematopoetic system we usually isolate cells from lymphoid organs, mainly spleen, thymus or bone marrow. Furthermore, to distinguish apoptosis responses of specific types of immune cells in culture, like different B- or T cell subsets derived from the spleen or, e.g. granulocytes from the bone marrow, cells are stained with specific cell surface markers selective for all myeloid cells MI/70, anti-Mac-1 (1:200); nucleated erythroid cells, Ter119 (1:50), T24.31.2, T cells, anti-Thy/CD90 (1:200); and granulocytes, Gr1-FITC (1:200) plus 7-AAD (1:100), to exclude dead cells. Cells are isolated using a FACS_Vantage cell sorter or FACS_ARIA_III (both BD).

Splenocytes negative for Mac-1 and Ter119 (Fig. 3A, dot-plot 3) are gated to a dot-plot representing T cells, which are positive for Thy1 (CD90) but negative for Gr1 (Fig. 3A, R4 on dot-plot 4). In this case the remaining Thy1−Gr1− cells represent mainly B cells in the spleen (Fig. 3A, R3 on dot-plot 4), as confirmed by post-sort B220 staining (not shown). In order to isolate granulocytes from the bone marrow we have set the gates simply on cells positive for Gr1-FITC (i.e. Gr1high) (Fig. 3B, on dot-plot 3). Sorted cells are washed (5 min, 600g at 4 °C) and resuspended in pre-warmed RPMI-1640 medium (PAA) supplemented with 250 μM l-Gln, 50 μM 2-mercaptoethanol, sodium pyruvate, nonessential amino acids (Invitrogen), penicillin/streptomycin (Sigma–Aldrich), and 10% FCS (PAA) at a density of 0.5–1.0 × 106/ml and cultured in an incubator at 37 °C and 5% CO2.

Fig. 3

Isolation of B and T cells from spleen and granulocytes from bone marrow. Viable cells were gated using 7-AAD staining (FL3-H) displayed in region 2 (R2) in dot-plots 2 (A and B) isolated from spleen and bone marrow. (A) Region 3 (R3) represent cells negative for Mac1-FITC and Ter119-FITC (dot-plot 3) which were gated on dot-plot 4 showing Thy1+Gr1− population representing splenic T cells (dot-plot 4, R4) and Thy1−Gr1− cells, i.e. B cells, were sorted from wt spleen (C57BL/6) (R3 in dot-plot 4). (B) Portion of bone marrow cells stained positive for Gr1-FITC (R3 in dot-plot 3) were sorted and confirmed to be >96% granulocytes (not shown). FSC (Forward scatter); SSC (Sideward scatter) is shown for each cell suspension in dot-plot 1.

Cells (5 × 104–1 × 105) are seeded into a 96-well flat-bottom plate in 200 μl media and left either untreated (Fig. 4A–C upper panel) or are stimulated with different apoptosis inducer, including staurosporine (100 nM), doxorubicin (400 ng/ml) or cisplatin (50 μM). After 12, 24 and 48 h of stimulation cells are harvested, washed in 1× Annexin-binding buffer (5 min, 600g at 4 °C) and co-stained with FITC-conjugated AnnexinV (1:1800 in Annexin-binding buffer, Biolegend; titration of each batch is recommend, as it can save a lot of money) and propidium iodide (5 μg/ml) for 5 min at 4 °C. Cell viability is monitored using a flow cytometer (FACSscan, Becton Dickinson) and analysed with WinMDIV2.8 Software. Sorted T cells from the spleen derived from wt mice on a mixed C57BL/6:129/SvJ (1:1) background were significantly more viable when compared to C57BL/6-wt, which was also observed after DNA-damage, induced by addition of doxorubicin or cisplatin (Fig. 4A). However, sorted B cells were equally viable independent of the genetic background (Fig. 4B). Interestingly, specific cell death in sorted granulocytes stimulated with cisplatin was increased when cells were derived from mixed C57BL/6:129/SvJ (1:1) genetic background (Fig. 4C). These results show clearly the importance of a uniform genetic background for cell death analysis in culture. Along similar lines, differences in sensitivity were observed in mouse embryonic fibroblasts (MEF) derived from E14.5 embryos of different backgrounds (Fig. 4D). After 1 h of heat-shock induction (43 and 45 °C) in an incubator, cells were cultured for additional 24 h at 37 °C before percentages of cell death were assessed by AnnexinV/PI co-staining. A minor but significant reduction of cell viability after heat-shock was observed in wt cells derived from mice of mixed background when compared to MEF on C57BL/6 background.

Fig. 4

Splenic T cells are more resistant against DNA damage on a mixed genetic background. Spontaneous cell death and specific cell death (%) give formula of sorted T- (A) B cells (B) and granulocytes from bone marrow (C) of C57BL/6 and mixed C57BL/6:129/SvJ mice over time. n > 3 per group, ∗p < 0.05. (D) Cellular viability after heat-shock treatment (1 h at 43 or 45 °C) of MEF derived from C57BL/6 and mixed C57BL/6:129/SvJ primary MEF (passage <3). After 24 h recovery at 37 °C viability was assessed by flow cytometry and propidium iodide exclusion. Means of n = 3 independent preparations are shown. ∗p < 0.05 compared to control treatment at 37 °C. Data were normalized using given formula (induced apoptosis − spontaneous cell death)/(100 − spontaneous cell death).

So, the overall take home message is: consider your genetic background and if you do not know it, specify, or have it specified for you, e.g. by PCR analysis of background-specific allelic markers (http://jaxmice.jax.org/geneticquality/index.html) [32]. This is of particular importance when immortalized MEF lacking individual genes are used and cells originate from different lab-sources or that may have been immortalized spontaneously or transformed using different oncogenes. Sometimes the background is simply not defined. Therefore significant differences observed during experiments may only be due to differences of background rather than to loss of the gene of interest. Even primary MEF in our hands show massive clonal variations; hence, wherever possible, genetic rescue experiments should be performed. Admittedly, a challenging and sometimes impossible task, e.g., when working with cell death inducing genes.

Monitoring apoptotic cell death in vivo

Whole body irradiation

In order to investigate the impact of pro-apoptotic proteins on DNA-damage induced by γ-irradiation in vivo we use whole body irradiation of mice at the age of 8–10 weeks. C57BL/6-wt, mixed C57BL/6:129/SvJ-wt and Casp2 (C57BL/6) mice are exposed to 1.75Gy of γ-irradiation. Tissues, i.e. bone marrow, thymus and spleen, were collected 20 h after irradiation. Single-cell suspensions are prepared and cells counted in a chamber hemocytometer (Neubauer). Interestingly, total cellularity of spleen and bone marrow was reduced after γ-irradiation derived from C57BL/6 mice, while cellularity was not significantly reduced in organs from C57BL/6:129/SvJ mice (Fig. 5A and B). Furthermore, no difference of organ cellularity was observed after γ-irradiation between wt and Casp2 mice on C57BL/6 background (Fig. 5B), indicating that caspase-2 has no impact on γ-irradiation induced cell death of lymphocytes in vivo while they seem even more vulnerable when compared to wt animals of mixed genetic background. In order to evaluate the effect of DNA-damage on the distribution of individual immune cell subsets in vivo, aliquots of the single cell suspensions were stained using fluorochrome-conjugated cell surface marker specific antibodies followed by flow cytometric analysis [33]. In Fig. 5C a comparison of splenic B cell subsets before and after whole body irradiation is depicted and demonstrates that B cell subsets are more sensitive to γ-irradiation-induced cell death when derived from C57BL/6 over C57BL/6:129/SvJ mice. Of note, this observation differs from our in vitro analysis where B cells of both genetic backgrounds died with similar kinetics, suggesting that the microenvironment also contributes to cell death rates in vivo. Such an effect is nicely documented in a study by Kelly et al. who showed that the relative resistance of Bad-deficient thymocytes depends on interaction with thymic stroma lacking Bad [22].

Fig. 5

Hematopoietic cells from C57BL/6:129/SvJ mice are more resistant to γ-irradiation in vivo. (A) Total cellularities (107) from untreated (control) and irradiated (1.75Gy) thymus, spleen and bone marrow from C57BL/6 and mixed C57BL/6:129/SvJ mice. ∗ Indicated differences refer to relevant controls. (B) As in (A) from wt (C57BL/6) and Casp2 mice (C57BL/6). (C) B cell subsets from C57BL/6 and mixed C57BL/6:129/SvJ mice before (untreated) and after whole body irradiation (1.75 Gy). §, ∗ indicate significant differences (p < 0.05) to relevant untreated controls. (D) Data are represent means ± SEM of >3 animal per genotype.

Conclusion

Investigating apoptotic cell death in vivo and in tissue culture is of critical importance to monitor the mechanisms involved in the maintenance of tissue homeostasis and the dynamics of developmental processes. In addition, new insights that are revealed in such studies can have important implications for pathologies such as tumorigenesis but also tumor therapy. While the analysis of cell lines has for sure advantages in cell biological as well as biochemical studies, especially since the discovery of RNA interference, certain limitations still apply that can be overcome by the usage of knockout mouse models or derived tissues and cell line systems. Nowadays, a large panel of different mouse models lacking individual cell death regulators are available to study these genes in normal physiology and disease. However, also there numerous pitfalls exist, that can limit the value of the data obtained in such studies as these tools often originate from different laboratories and are frequently not of uniform genetic background, compromising results and comparability. Hence, it is crucial to choose a uniform genetic background for such assays and if cells or cell lines used are of uncertain genetic background origin, proper genotyping and rescue experiments must be considered compulsory.