Mycoplasma Infection Mediates Sensitivity of Multidrug-Resistant Cell Lines to Tiopronin: A Cautionary Tale

Abstract

We previously reported that some, but not all, multidrug-resistant cells that overexpressed various drug-resistance transporters were collaterally sensitive to tiopronin. In recent follow-up studies, we discovered that sensitivity to tiopronin in the original study was mediated by infection of the cells by a human-specific strain of mycoplasma. These results strongly support the need to constantly monitor cells for mycoplasma infection and keep stored samples of all cells that are used for in vitro studies.

INTRODUCTION

The ATP-binding cassette transporters P-glycoprotein (P-gp, encoded by the ABCB1 gene), multidrug resistance-associated protein 1 (MRP1, encoded by ABCC1) and the breast cancer resistance protein (BCRP or ABCG2, encoded by the ABCG2 gene) are known to confer multidrug resistance (MDR) or cross-resistance to multiple, structurally different compounds.¹ Collateral sensitivity has emerged as a potential way to target multidrug-resistant cells, whereby agents are found that are more toxic to the resistant cells versus cells that are not drug-resistant. For example, verapamil has been found to be more toxic to cells that overexpress P-gp ² or MRP1.³ The extensive research on collateral sensitivity has been reviewed in Szakacs et al.⁴ and Pluchino et al.⁵ We previously reported that the thiosemicarbazone NSC73306 is selectively toxic to cells that overexpress P-gp, but not MRP1 or ABCG2.⁶ Separately, we unexpectedly found that the orphan drug tiopronin was considerably more toxic to some cells that expressed P-gp or MRP1.⁷ We thus sought to explore tiopronin as a potential treatment for cancers that expressed high levels of transporters associated with drug resistance.^{7, 8} Although high concentrations of tiopronin were needed to elicit collateral sensitivity, we attempted to synthesize analogs with higher potency in the hope of finding a lead compound suitable for clinical evaluation.⁷ The observed apoptosis appeared to be mediated by reactive oxygen species (ROS) generated by inhibition of glutathione peroxidases 1 and 4.⁸

As part of determining the molecular basis of this selective sensitivity to tiopronin, we thawed frozen stocks of drug-resistant lines that we previously reported as sensitive to tiopronin and found that they were no longer exquisitely sensitive to the drug. When we thawed cells that were frozen while we were preparing our initial publication (approximately 2011), we found that the cells were still sensitive, suggesting an unknown environmental perturbation. Indeed, we originally suspected that some unknown factor might be responsible for the variable sensitivity to tiopronin among the cell lines studied, as the first sentence of the discussion of Goldsborough et al. ⁷ states: “We have shown here that while not all MDR cell lines are collaterally sensitive to 1, its cytotoxic activity is independent of P-gp or MRP1 drug efflux transporters and may therefore be targeting a more general feature of MDR cells that has been acquired during drug selection.”

Upon further characterization of the cell lines, we were able to determine that the drug-selected cell lines that were previously sensitive to tiopronin were also infected with mycoplasma. When drug-resistant cell lines that were not sensitive to tiopronin were exposed to mycoplasma, sensitivity was restored. Similarly, when tiopronin-sensitive, drug-resistant cell lines were cured of mycoplasma with antibiotics, they became once again resistant to tiopronin treatment. We find that infection of drug-resistant cell lines with mycoplasma sensitized cells to tiopronin, but not parental, drug-sensitive cell lines, most likely via ROS induction. We thus provide evidence that mycoplasma infection can increase the sensitivity of multidrug-resistant cell lines to some drugs.

RESULTS

Mycoplasma Infection Determines Response of Multidrug-Resistant Cell Lines to Tiopronin.

Figure 1A details the various uninfected and mycoplasma-infected KB-3-1 and KB-V1 cells used during the course of the study. We previously demonstrated that the vinblastine-selected cell line, KB-V1, was approximately 50-fold more sensitive to the drug tiopronin (chemical structure shown in Figure 1B) than the parental cell line KB-3-1.⁷ However, when recently attempting to replicate the data, the cells were no longer found to be collaterally sensitive to the drug (Figure 2A). When examining cells that had been cryopreserved about the same time that the previous publication was generated (2011), we found that the MDR cells were indeed sensitive to tiopronin. When the cells were checked for mycoplasma contamination, the KB-V1 cell line that was sensitive to tiopronin was found to be mycoplasma infected (termed KB-V1 (+)C). To determine if the mycoplasma are responsible for the observed sensitivity to tiopronin, cells were exposed to Mycoplasma Removal Agent according to the manufacturer’s instructions until the cells were found to be ‘cured’ of mycoplasma (termed KB-V1 (+)C/MRA). Of course, it must be understood that removal of mycoplasma would not necessarily reverse any potential cellular damage caused by the initial infection. Nonetheless, when cytotoxicity assays were performed with tiopronin on the cells that had been cured of mycoplasma, the cells were no longer sensitive (Figure 2B). Similarly, we took the KB-V1 cells that did not have mycoplasma infection (KB-V1 (−)) and infected them with mycoplasma by adding filtered medium from infected cells (termed KB-V1 (+)A). Once mycoplasma infection was confirmed, we again performed cytotoxicity assays with tiopronin and found that the previously insensitive cells had become sensitive to tiopronin (Figure 2C). Once the acutely infected KB-V1 (+) A cells were treated with MRA to remove mycoplasma infection (termed KB-V1 (+) A/MRA), sensitivity to tiopronin was diminished. We also took uninfected parental cells (KB-3-1 (−)) and infected them with mycoplasma by exposing cells to filtered medium from infected cells. Mycoplasma infection did not considerably change the sensitivity of KB-3-1 cells to tiopronin (Figure 2D). Cytotoxicity assay results are summarized in Tables 1 and 2. Upon typing the mycoplasma from the KB cell lines to determine their origin, they were found to be Mycoplasma fermentans (IDEXX BioAnalytics, Westbrook, ME).

Figure 1:

Treatment schema and chemical structure (A) Various uninfected and mycoplasma-infected KB-3-1 and KB-V1 cells were studied; (B) Chemical structure of tiopronin.

Figure 2:

Mycoplasma infected MDR cells are sensitive to tiopronin treatment. Representative three-day cytotoxicity results with tiopronin on: (A) mycoplasma-free KB-3-1 (−) and KB-V1 (−) cells; (B) mycoplasma-free KB-3-1 (−) and KB-V1 (−) cells as well as chronically infected KB-V1 (+) C cells or chronically infected cells cleared of mycoplasma, KB-V1 (+) C/MRA; (C) mycoplasma-free KB-3-1 (−) and KB-V1 (−) cells as well as acutely infected KB-V1 (+) A cells or acutely infected cells cleared of mycoplasma, KB-V1(+)A/MRA; (D) mycoplasma-free KB-3-1 (−) and acutely infected KB-3-1 (+) cells.

Table 1.

Summary of cytotoxicity data on KB-3-1 cells with and without mycoplasma infection^a

Drug	KB-3-1	KB-3-1 A (+)	KB-3-1 A (+) MRA
Tiopronin (mM)	10.1 (+/−3.8)	6.8 (+/− 1.8)	7.1 (+/− 1.2)
Doxorubicin (µM)	.09 (+/− .07)	.11 (+/− .06)	.07 (+/− .06)
Vincristine (µM)	.0003 (+/− .00017)	.00043 (+/− .00012)	.00029 (+/− .00011)

^aResults presented are average IC₅₀ values +/− standard deviation (SD).

Table 2.

Summary of cytotoxicity data on KB-V1 cells with and without mycoplasma infection^a

Drug	KB-V1	KB-V1 C (+)	KB-V1 C (+) MRA	KB-V1 A (+)	KB-V1 A (+) MRA
Tiopronin (mM)	7.7 (+/− 2.3)	.27 (+/−.12)	7.0 (+/− 3.0)	.36 (+/− .11)	7.33 (+/− 2.5)
Doxorubicin (µM)	6.07 (+/− 4.56)	10.67 (+/− 4.04)	9.0 (+/− 2.0)	11.33 (+/− 2.3)	7.4 (+/− 3.9)
Vincristine (µM)	1.08 (+/− .34)	2.38 (+/− 1.83)	1.30 (+/− .58)	2.05 (+/− .38)	1.35 (+/− .53)

^aResults presented are average IC₅₀ values +/− standard deviation (SD).

Mycoplasma Infection Does Not Affect Expression of P-gp or Drug Sensitivity.

We next determined if mycoplasma infection had any effect on expression of P-gp or drug sensitivity in the KB-V1 cells. Expression of P-gp was measured by flow cytometry using the specific antibody UIC2. Parental KB-3-1 cells do not express P-gp (Figure 3A, left histogram) as shown by the overlapping blue and red histograms. Expression of P-gp as measured by UIC2 antibody staining was high, as shown by the distance between the red and blue histograms and didnot change in KB-V1 (−) cells (Figure 3A, second histogram) compared to chronically infected KB-V1 (+) C (third histogram) or chronicallyinfected KB-V1 (+) C/MRA cells that were free of mycoplasma (last histogram). P-gp expression was also determined by immunoblotting with the anti-Pgp antibody C219. Comparable levels of P-gp were found in mycoplasma-infected and mycoplasma-free KB-V1 cell lines; no Pgp was detected in KB-3-1 parental cells (Figure S1). Sensitivity to the P-gp substrates doxorubicin (DOX) and vincristine (VCR) was then assessed by 3-day cytotoxicity assays. As shown in Figure 3C, mycoplasma infection did not appreciably change sensitivity to DOX or VCR. Results are summarized in Tables 1 and 2.

Figure 3:

Mycoplasma infection does not affect expression or function of P-gp. (A) Mycoplasma-free KB-3-1 (−) and KB-V1 (−) cells as well as chronically infected KB-V1 (+) C cells or chronically infected cells cleared of mycoplasma, KB-V1 (+) C/MRA were incubated with phycoerythrin-labeled isotype control antibody (red histograms) or phycoerythrin-labeled UIC2 antibody (blue histograms) for 20 min in 2% BSA after which cells were washed with PBS and examined by flow cytometry. Three-day cytotoxicity assays were performed with doxorubicin (B), vincristine (C) or NSC73306 (D) on the cells as noted in Figure 1A. (E) Three-day cytotoxicity assays were performed with tiopronin on mycoplasma-free KB-3-1 (−) and KB-V1 (−) cells as well as chronically mycoplasma infected KB-V1 (+) C cells in the presence or absence of 5 mM NAC.

We previously reported that cells that overexpress P-gp are collaterally sensitive to NSC73306⁶, leading us to evaluate whether mycoplasma infection might also affect sensitivity to this compound. As shown in Figure 3D, mycoplasma infection did not affect sensitivity of P-gp overexpressing KB-V1 cells to NSC73306.

Cell Death Mediated by Mycoplasma Infection and Tiopronin in P-gp Positive Cells Can Be Reversed by N-acetylcysteine.

Since we previously found that the toxicity of tiopronin to cells that overexpressed transporter proteins appeared to be mediated by inhibition of glutathione peroxidases 1 (GPx1) and 4 (GPx4) leading to the generation of ROS⁸, we performed cytotoxicity assays on chronic mycoplasma infected KB-V1 cells in the presence or absence of 5 mM N-acetylcysteine (NAC). As can be observed in Figure 3E, the sensitivity of the chronically infected KB-V1 (+) C cells was reversed by the addition of NAC. Interestingly, when we thawed the KB-V1 cells that were treated with shRNAs to knock down GPx1 or GPx4, we found that these were also mycoplasma infected. Thus, our previous findings suggesting that the increased sensitivity of transporter positive cell lines to tiopronin due to ROS induction mediated by GPx1 and GPx4⁸ appear to be true for mycoplasma infected cells.

DISCUSSION

Mycoplasma are single-celled prokaryotes that are devoid of a cell wall, and the primary origin of mycoplasma in continuous cell culture is believed to originate from the human oral cavity (i.e. laboratory cell culture operators, speaking and breathing).⁹ The contamination of in vitro cell culture by mycoplasma was first reported in 1956.¹⁰ Since that time, individual reports of the effect of mycoplasma contamination on cellular phenotypes have appeared in the literature, though few examples of systematic studies of the impact of mycoplasma have been reported.¹¹ Over the past two decades, there has been an increasing awareness of the impact of mycoplasma contamination on reproducibility in the scientific literature, and screening of cell culture collections have produced estimates that 15–35% of continuous cell cultures are mycoplasma contaminated ^{12, 13} (and much higher numbers have been reported in individual collections). These findings have stimulated the development of off-the-shelf mycoplasma detection kits for routine laboratory screening for mycoplasma contamination.

While many labs do not notice mycoplasma contamination specifically because there are no universal phenotypes associated with contamination, a number of research-related phenotypes have been reported to be due to mycoplasma contamination rather than the underlying biology alone. In the context of cancer experimental therapeutics, mycoplasma infection has previously been shown to affect cancer cell line response to chemotherapy. For example, Liu and colleagues noted increased sensitivity to cisplatin, gemcitabine and mitoxantrone in HCC97L hepatocarcinoma cells infected with Mycoplasma hyorhinis compared to uninfected cells.¹⁴ Consistent with our findings, they reported that mycoplasma infection did not affect levels or localization of ABCB1, ABCC1 or ABCG2. Mycoplasma hyorhinis was also found to decrease sensitivity to the MDM2 inhibitor Nutlin-3 in NCI-H292 lung cancer cells.¹⁵ Mycoplasma contaminated HCT-116 colon cancer cells were found to be 5- to 100-fold more resistant to 5-fluorouracil and 5-fluorodeoxyuridine, respectively, compared to cells that were cured of infection with antibiotics.¹⁶

In cases where genuine scientific error is identified, we believe correcting the literature is essential to ensure reproducibility. While the exact mechanism by which tiopronin is selectively toxic to mycoplasma-infected multidrug-resistant cell lines is unclear, it is robust and appears to be due to ROS-induced stress, as knockdown of GPx1 and GPx4 were shown to modulate cell death and NAC was able to prevent tiopronin-mediated cell death. As cells with knockdown of GPx1 and GPx4 were also found to be contaminated with mycoplasma, this mechanism appears to be applicable here. The data in our original reports has been reproduced, but with the new understanding that the presence of mycoplasma contamination was necessary for the sensitivity to tiopronin.

Given the high (>15%) tissue culture mycoplasma contamination rates reported in the literature, it is likely that a significant number of published phenotypes may be due to mycoplasma contamination. Along with routine mycoplasma testing, to ensure that findings are robust, we suggest that phenotypes studied in long-term laboratory cell cultures be re-confirmed using the same cell line re-sourced from a cell line repository or originating laboratory. Although we routinely check our stock cell cultures for the presence of Mycoplasma, our findings reported here underscore the importance of regularly screening for mycoplasma infection in all cell lines being cultivated for experiments in the laboratory and serves as a cautionary tale to those performing cytotoxicity assays or examining cell death.

EXPERIMENTAL SECTION

Cell Lines.

KB-3-1 and its P-gp-expressing subline KB-V1 (maintained in 1 µg/ml vinblastine sulfate), were cultured in RPMI supplemented with 10% FBS, 2 mM glutamine, 100 units/ml penicillin and 100 µg/ml streptomycin.

Chemicals.

Tiopronin, N-Acetylcysteine, vincristine sulfate, and paclitaxel were obtained from Sigma Chemical, (St. Louis, MO). Mycoplasma Removal Agent (MRA) was obtained from BioRad, (Hercules, CA). Cell Titer Glo reagent was purchased from Promega Corporation (Madison, WI).

Flow Cytometry.

P-gp expression was determined using phycoerythrin-labeled UIC2 antibody paired with an appropriate phycoerythrin-labeled isotype control (both from Thermo-Fisher, Waltham, MA). Cells were trypsinized and incubated with antibodies in 2% BSA for 20 minutes at room temperature after which cells were washed twice with phosphate-buffered saline (PBS) and analyzed by flow cytometry. A FACSCanto II flow cytometer (Becton Dickinson Bioscience, San Jose, CA) equipped with a 488-nm argon laser and a 585-nm bandpass filter was used to detect phycoerythrin fluorescence. At least 10,000 events were collected for each sample.

Cytotoxicity Assays.

Cells were seeded into 96-well plates at a concentration of 5000 cells per well in cell culture media and were allowed to attach overnight. Cytotoxic agents were subsequently added in serial dilutions. Cell viability was measured 72 hours after drug was added using the Cell Titer Glo luminescence-based detection kit (Promega, Madison, WI)). Cytotoxicity (IC₅₀) was defined as the drug concentration that reduced cell viability to 50% of the untreated control cells. Collateral sensitivity of the drug-resistant cell line was calculated by dividing the IC₅₀ of the parental cell line by the IC₅₀ of its drug-resistant derivative.

Mycoplasma Testing and Treatment.

Cell lines in the laboratory were checked using the MycoAlert PLUS Assay kit from Lonza (Walkersville, MD). For our studies, we treated mycoplasma positive cell lines with the Mycoplasma Removal Agent (MRA) (a broad-spectrum antibiotic containing a 4-oxoquinoline-3 carboxylic acid derivative) at a concentration of 0.5 µg/ml in a 6-well plate. Media + MRA was removed and replaced every 48 hr. Cells were re-tested for mycoplasma at 7 days after the end of the treatment period.

ABBREVIATIONS

P-gp

P-glycoprotein

MRP1

multidrug resistance-associated protein 1

ROS

reactive oxygen species

MDR

multidrug resistance

MRA

Mycoplasma Removal Agent

NAC

N-acetylcysteine

PBS

phosphate-buffered saline