Sophorolipid Candidates Demonstrate Cytotoxic Efficacy against 2D and 3D Breast Cancer Models

Abstract

Sophorolipids are biosurfactants derived from the nonpathogenic yeasts such as Starmerella bombicola with potential efficacy in anticancer applications. Simple and cost-effective synthesis of these drugs makes them a promising alternative to traditional chemotherapeutics, pending their success in preliminary drug-screening. Drug-screening typically utilizes 2D cell monolayers due to their simplicity and ease of high-throughput assessment. However, 2D assays fail to capture the complexity and 3D context of the tumor microenvironment and have consequently been implicated in the high percentage of drugs investigated in vitro that later fail in clinical trials. Herein, we screened two sophorolipid candidates and a clinically-used chemotherapeutic, doxorubicin, on in vitro breast cancer models ranging from 2D monolayers to 3D spheroids, employing optical coherence tomography to confirm these morphologies. We calculated corresponding IC₅₀ values for these drugs and found one of the sophorolipids to have comparable toxicities to the chemotherapeutic control. Our findings show increased drug resistance associated with model dimensionality, such that all drugs tested showed that 3D spheroids exhibited higher IC₅₀ values than their 2D counterparts. These findings demonstrate promising preliminary data to support the use of sophorolipids as a more affordable alternative to traditional clinical interventions and demonstrate the importance of 3D tumor models in assessing drug response.

Figure 1.

Representative day 4 OCT images showing development of MDA-MB-231 aggregates. The cellular aggregates (darker regions) are visible atop the plastic wells (lighter regions) in which they are grown. While both samples appear as spheroids in 2D en face view, only those samples with Matrigel (column B) formed spheroids. (A) Matrigel (−) samples formed loose, flattened, disk-like aggregates with limited compaction. Scale bars = 100 μm.

Figure 2.

Quantification of day 4 aggregate volume and sphericity in the MDA-MB-231 cell line showed significant differences in aggregate shape despite no differences in volume. A cutoff sphericity of 0.8 was used to discern tumor spheroids from nonspherical tumor aggregates. Matrigel (+) samples produced highly spherical MCTSs when compared with Matrigel (−) samples, which formed flatter, disk-like aggregates (n = 6/condition) (** = p < 0.01).

Figure 3.

IC₅₀ values obtained across conditions revealed increased drug resistance in 2.5D and 3D models compared to 2D monolayers. Both SL compounds exhibited a large decrease in cytotoxicity from 2D monolayer values to those found in 2.5D and 3D aggregate models. However, the addition of Matrigel did not cause a further reduction in toxicity on MDA-MB-231 from 2.5D to 3D aggregates, as previously hypothesized (n = 3/concentration/trial, 3 independent trials performed for each morphology and each drug*; *two independent trials performed for 2.5D DLSL).

Figure 4.

Longitudinal Imaris cell counts in MDA-MB-231 aggregates before and after 12 h SL treatment at respective IC₅₀ doses, presented alongside dissociated counts from the same aggregates. The number of cells present decreases for both 2.5D and 3D aggregates after 12 h treatment with DLSL (A, C) and SL-EE-D (B, D). Dissociated cell suspensions obtained from the same aggregates postimaging and stained with trypan blue revealed a large proportion of necrotic cells (A–D, right side). Untreated MDA-MB-231 OCT aggregate data (E, F, left side) show no decrease in cell count over the 12 h period, while trypan blue staining registered necrotic cells within the models (E, F, right side). This necrosis is likely due to a lack of media and oxygen diffusion into the center of the aggregate and is more pronounced in 3D aggregates (n = 3 aggregates, HC n = 1 aggregate).

Figure 5.

GI₅₀ determination for MDA-MB-231 models treated with doxorubicin revealed an increase in drug resistance associated with increasing model dimensionality (from monolayer to aggregate/spheroid) (n = 3/concentration).

Figure 6.

IC₅₀ curves for SL-EE-D and DLSL on human dermal fibroblasts in a traditional 2D monolayer. Values were 1.5–2× higher than those found for MDA-MB-231 cells, indicating that the SLs possess minor cancer cell selectivity (n = 3/concentration/trial, 3 independent trials performed).

Figure 7.

HC₅₀ determinations for DLSL and SL-EE-D. Approximately 50% red blood cell lysis occurred at 125 μg/mL for both compounds. Data were collected in triplicate.

Figure 8.

Structures of natural and modified sophorolipids studied herein: (A) natural lactonic sophorolipid diacetylated at the sophorose 6′ and 6″ ring positions (DLSL). (B) Sophorolipid ethyl ester diacetylated at the 6′ and 6′′ sophorose ring positions (SL-EE-D).