Optical study of chemotherapy efficiency in cancer treatment via intracellular structural disorder analysis using partial wave spectroscopy

Abstract

As cancer progresses, macromolecules, such as DNA, RNA or lipids, inside cells undergo spatial structural rearrangements and alterations. Mesoscopic light transport-based optical partial wave spectroscopy (PWS) was recently introduced to quantify changes in the nanoscale structural disorder in biological cells. The PWS measurement is performed using a parameter termed as “disorder strength” ($L_d$), which represents the degree of nanoscale structural disorder inside the cells. It was shown that cancerous cells have higher disorder strength than normal cells. In this work, we first used the PWS to analyze the hierarchy of different types of prostate cancer cells, namely, C4–2, DU-145 and PC-3, by quantifying their average disorder strengths. Results expectedly showed that $L_d$ values increases in accordance with the increasing aggressiveness/tumorigenicity levels of these cells. Using the $L_d$ parameter, we then analyzed the chemoresistance properties of these prostate cancer cells to docetaxel drug compared to their chemosensitivity. Results show that chemoresistant cancer cells have increased $L_d$ values, that is, higher disorder strength, relative to chemosensitive cancer cells. Thus, use of the $L_d$ metric can be effective in determining the efficacy of particular chemotherapy.

Graphical Abstract

1 |. INTRODUCTION

1.1 |. Role of nanoscale mass density fluctuations in cancer detection

It is well known that the progress of cancer is associated with cellular morphological changes [1, 2]. Based on mesoscopic light transport theory, a biophotonics technique, namely, optical partial wave spectroscopic (PWS) microscopy was introduced earlier to quantify the nanoscale structural disorder in weakly disordered optical systems such as biological cells [3, 4]. Since then, PWS has been used to study different cancerous and noncancerous cells by quantifying their nanoscale intracellular refractive index fluctuations in term of their disorder strength [3–8]. As a test, in this work, we used PWS technique to analyze structural disorder properties of three prostate cancer cells, namely, C4–2, DU-145 and PC-3, and the effect of the treatment of chemotherapy drug, docetaxel, to these cancer cells.

PWS analysis is based on statistical quantification of backscattered intensities and their spectral correlation decay length owing to the nanoscale refractive index fluctuation within biological samples [3, 4]. In the PWS technique, a thin sample is virtually divided into several parallel scattering subsamples such that the bulk backscattering problem is approximated as a quasi-one-dimensional (1D) scattering problem [3, 4]. The reflection in each 1D channel, that is, subsample, provides two parameters: root mean square (RMS) and the correlation of the reflection spectra. From these two quantities, the effective structural disorder, or the “disorder strength $L_d$,” inside the biological cell is characterized. The parameter $L_d$ is defined as, $L_d=<d{n}^2>\times l_c$, where $dn$ is the local fluctuating part of the refractive index in the 1D channel, arising from the mass density fluctuations in that channel, and $l_c$ is its spatial correlation length [3, 4]. Consequently, $L_d$ represents a measure of refractive index fluctuation, and hence mass density fluctuation, in the sample, both in terms of magnitude and correlation length. This study used disorder strength, $L_d$, as the primary metric to determine the efficacy of docetaxel, our model chemotherapeutic drug, against prostate cancer cell lines. In particular, we demonstrated that the intracellular structural properties of drug-resistant vs drug-sensitive cells show different $L_d$ values, strongly suggesting that the sensitivity or resistance of cancer cells to a given drug can be evaluated using the computed disorder strength parameter $L_d$. As such, the parameter $L_d$ acts as a potential biomarker of drug efficacy, as will be shown in our results. Briefly, drug-resistant cancer cells have higher disorder strengths, compared to that of normal (control) cancer cells, suggesting higher aggressiveness of the drug resistant cancer cells. For that, we first show, by analyzing three different prostate cancer cells (namely, C4–2, DU-145 and PC-3), that higher aggressiveness/tumorigenicity levels of cancer cell results in higher disorder strengths in those cells, which is consistent with earlier reported results with other cancer cases [3–9]. Subsequently, the drug treatment study of these cells was performed. These results indicate that drug-sensitive and drug-resistant cancer cells display quantifiable differences in their intracellular structural disorder. These nanoscale alterations are, in turn, related to nanoscale mass-density fluctuations inside the cells. Therefore, the degree of disorder strength, $L_d$, can be considered as a potential biomarker, whose values can be used to assign relative efficacy to a given chemotherapy drug treatment, that is, a marker that can differentiate between drug resistance and drug sensitivity. This will be demonstrated in the following sections.

1.2 |. Prostate cancer and effect of chemotherapy drug

Prostate cancer is most commonly diagnosed among men and often metastasizes at later ages. An alarming total of 164 690 new cases are estimated in the United States in 2018, including 29 430 deaths [10]. About one in seven men will be diagnosed with prostate cancer during their lifetime. Prostate cancer is prominent in older men, with 60% of men over the age of 65 are diagnosed with this disease. Prostate cancer is second leading cause of cancer caused deaths in American men, behind only to the lung cancer. Therefore, effective detection and treatment of prostate cancer are important, especially for men of advanced age. Chemotherapy is used to treat metastasized prostate cancer. However, chemotherapy is often ineffective because cells of individual patients' tumors develop chemoresistance [11–13]. Thus, it is important to know the efficacy of a particular chemotherapeutic drug in order to assess its potential resistance or, conversely, its potential sensitivity.

1.3 |. Prostate cancer treatment and role of chemotherapy

Prostate cancer treatment typically involves radiation, surgery, and then chemotherapy. These treatments are performed in an order that matches the stage and size of the malignant tumor. Usually, chemotherapy is administered after surgery to remove the tumor and prostate. This therapy is expected to kill the metastasized cancer cells around the prostate gland, as well as surrounding tissue. Resistance to chemotherapy is one of the main causes of treatment failure in all types of cancer treatment, including that of prostate cancer. Development of drug resistance by cancer cells is a serious problem since it ends the remission stage and leads to disease relapse. Clinically, the chemotherapy drug docetaxel (Taxotere) is frequently used to treat the advanced stages (metastasis) of prostate cancer. However, resistance to this drug is a major clinical problem as it was established as the primary drug therapy for metastatic prostate cancer [13]. Docetaxel resistance is correlated with treatment time and drug dosage [14]. Researchers are studying drug resistance with the aim of understanding the physiological effect of other chemotherapeutic drugs that can bypass resistance and restore near-normal growth and division processes for cells that remain after chemotherapy [15–17]. In different chemoresistance studies, the aggressiveness and invasion of resistant cancer cells have been observed [18, 19]. In addition, it has been reported that docetaxel resistance in cancer cells is highly associated with genetic alterations [20]. These genetic mutations lead to the rearrangement and alteration of the most basic building blocks of cells, such as DNA, RNA and lipids, in turn producing intracellular structural changes in the treated cancer cells [20]. In this work, the effect of this drug on intracellular structural changes was measured and quantified using the parameter disorder strength, $L_d\left(=<d{n}^2>\times lc\right)$, as defined earlier. The resulting value was compared to the corresponding values for the nondrug-treated, that is, the drug sensitive cancer cells. The results support the hypothesis that prostate cancer cells that acquire chemoresistance to docetaxel, ultimately showing no response toward further treatment with this drug, became more aggressive. Consequently, we have demonstrated that $L_d$ can potentially be an efficient metric (/biomarker) that can characterize drug-sensitive vs drug-resistant cancer cells using PWS technique.

2 |. METHODS

2.1 |. Cell culture and development of docetaxel chemoresistant prostate cancer cells of the following cell lines: C4–2, DU145 and PC-3

Prostate cancer cell lines (C4–2, DU-145 and PC-3) were developed in cell culture facilities at the University of Tennessee Health Science Center (UTHSC) and cultured in RPMI-1640 medium containing 10% (vol/vol) fetal bovine serum, 2 mM l-glutamine (Invitrogen, Carlsbad, California) and 1% (wt/vol) penicillin-streptomycin (Gibco; Thermo Fisher Scientific, Grand Island, New York) at 37C in a humidified 5% CO₂-95% air atmosphere (Thermo Fisher Scientific, Waltham, Massachusetts). Prostate cancer-resistant lines were generated by initial treatments with docetaxel (MP Biomedicals, Santa Ana, California) at 1 nM (from 5 μM stock) in 75 cm² flasks for 24 to 48 hours. After treatment, the surviving cells were reseeded into new flasks and allowed to recover for 1 to 2 days. Cells were maintained at 1 nM up to four treatment cycles (4 TC). Gradually, the concentration of docetaxel was increased to 2.5 nM (6 TC). The 5 and 10 nM of docetaxel was then continued to 8 and 12 TC, respectively. Then all cells underwent 12 TC and 15 TC with docetaxel 15 and 20 nM, respectively. Finally, at 30 nM docetaxel, we performed 25 TC to ensure the acquisition of chemoresistance in all C4–2, D-145 and PC-3 cells. Following each treatment, cells were allowed to fully recover before assessing their resistance to docetaxel and performing any experimental work. In total, the cells were treated for eight long months. Since the passage number of the drug-treated cells increased over time, a subset of prostate cancer cells was aged alongside these cells as an appropriate control to ensure that the effects seen resulted from resistance and not the aging effect of the prostate cancer cells.

2.2 |. Cell imaging and analysis

Frozen cell batches with similar passages were thawed and used for all experiments. For image analysis studies, we seeded 2.5 × 10⁴ cells in each well of four-chambered slides (Sarstedt Inc, Newton, North Carolina) and allowed them to grow. After reaching 70% to 80% of confluence, cells were fixed using 4% paraformaldehyde for 20 minutes [21, 22]. After incubation, cells were washed with PBS, followed by PWS imaging and analysis.

3 |. OPTICAL PWS EXPERIMENT

3.1 |. Instrumentation

A schematic of the experimental apparatus used for PWS is shown in Figure 1. A broadband white light from a Xenon Lamp (Newport Corp., Irvine, California) is used to illuminate the sample, using Kohler illumination.

FIGURE 1.

Experimental setup for PWS

First, the light beam from the source is directed through a broadband dielectric mirror (R > 99%), and then is passed through a set of lenses and apertures to collimate it, as shown in Figure 1. The collimated light beam is redirected by a right angle prism, allowing it to pass through a beam splitter plate (BSP, 50:50) to a low numerical aperture objective (NA = 0.65, ×40). Next, the beam is focused on the sample, which is kept on an electronic motorized stage (Zaber Technology Inc., Vancouver, Canada) with precise resolutions around 100 nm in vertical direction ($z$-axis) and 40 nm in $x-y$ plane such that the depth of the image is focused more accurately and in controlled manner than permitted in previous versions [3, 4]. The backscattered light passes through the objective again and is projected into a visible range (450–700 nm) liquid crystal tunable filter (Varispec; PerkinElmer Inc., Waltham, Massachusetts) where the signal is filtered according to its wavelength components. The filtered signal is captured by the CCD camera detector (CoolSNAP; Photometrics, Tucson, Arizona), and the image of the sample is observed on the computer screen.

3.2 |. ${\mathit{L}}_{\mathit{d}}$ value from backscattering intensity

The CCD camera captures the backscattered image of the biological cell taken at different wavelengths ($\lambda$) in the visible range, 450– to 700 nm. Eventually, a data cube of backscattered intensity $I(x,y;\lambda )$, where ($x,y$) represents the spatial position on the sample, are acquired by the PWS system. For further data processing, we extract the backscattered spectral fluctuation $R(x,y;\lambda )$ as described elsewhere [3, 4]. Subsequently, the disorder strength $L_d$ is derived from two quantities: RMS of the reflection intensity ${<R>}_{\text{rms}}$ and spectral correlation decay of the reflection intensity $C\left(\Delta k\right)$, as follows [3, 4],

$$L_d={\left.\frac{B⟨R{⟩}_{\mathrm{r}\mathrm{m}\mathrm{s}}}{2k^2}\frac{(\mathrm{\Delta }k{)}^2}{-\mathrm{l}\mathrm{n}\left(C\right(\mathrm{\Delta }k\left)\right)}\right|}_{\mathrm{\Delta }k\to 0},$$

where $B$ is the normalization constant, and $C\left(\Delta k\right)$ is the autocorrelation function of $R(x,y,k)$ at a particular position ($\mathrm{x},\mathrm{y}$) averaged over many ensembles. The average value of $L_d$ (averaged over ensemble) is calculated for each pixel point ($x,y$). As pointed out earlier, the $L_d$ value measured at any spatial point ($\mathrm{x},\mathrm{y}$) represents a measure of refractive index fluctuation at that point along the depth of the sample ($L_d=<d{n}^2>\times l_c$).

3.3 |. Statistical analysis

One-way ANOVA was used to test overall significant difference among the groups. When ANOVA detects a significant difference among all the different groups studied, the differences between the pair groups were examined using Tukey's post hoc test. All statistical analyses were performed using Sigma Stat software (SPSS Inc., Chicago, Illinois). The significant level for all groups was set at $\alpha =0.05$.

4 |. RESULTS AND DISCUSSION

The disorder strength measurement was first performed on the three prostate cancer cell lines, namely, C4–2, DU-145 and PC-3, for which the hierarchy of tumorigenicity (aggressiveness or metastatic potential) have been well characterized [23–25]. The measurements were conducted on ~20 cells randomly selected from each cell type in one set of measurement (~60 cells in three sets). The results are shown in Figure 2.

FIGURE 2.

Representative bright-field images (A-C) and corresponding two-dimensional (2D) $L_d$ images (A’-C’) (PWS images) for the three prostate cell lines, DU-145, C4–2 and PC-3, respectively. The scale bar on all the images represents 10 μm

In Figure 2A–C, three representative bright-field images of each cell lines, namely, DU-145, C4–2 and PC-3 are presented, and their corresponding PWS $L_d$ images are shown in Figure 2A’–C’, respectively. The $L_d$ color map shows the distribution of $L_d$ values in the cells. In $L_d$ color map, the disorder strength increases from blue color to red. As pointed out earlier, a higher $L_d$ value at any pixel point ($\mathrm{x},\mathrm{y}$) in the $L_d$ image indicates higher degree of structural disorder along the depth of the cell at that particular spatial point. It can be seen in Figure 2A’–C’ that the $L_d$ distribution for the three cell lines are distinctly different. Subsequently, therefore, we calculated the average disorder strength, that is, mean $L_d$ value, of each cell lines by taking the ensemble average of $L_d$ values of all the cells in three different sets (n = 60). The results are shown in Figure 3. The bar graph in Figure 3 shows a comparison of the mean $L_d$ values of the DU-145, C4–2 and PC-3 cell lines. As it is clear from Figure 3, the disorder strengths in these cell lines are in the order of DU-145 < C4–2 < PC-3, which is in accordance with their metastatic potential level (aggressiveness/tumorigenicity) as well [23–25]. Consequently, the present result suggests that the $L_d$ value (ie, the disorder strength) is directly correlated with the aggressiveness level of the cell lines.

FIGURE 3.

Bar plots of mean intracellular disorder strength $L_d$ calculated for the three types of human prostate cancer cell lines: DU-145, C4–2 and PC-3. $L_d$ results show that the average $L_d$ value correlates with the tumorigenicity level of the cell. Statistically, there was significant difference between the three groups, F_(3,20) = 15.351 (P < 0.05). Tukey's post hoc test was used to determine the significance between each pair of the three groups (P < 0.05), PC3 has the greatest mean intracellular disorder strength and DU-145 has the lowest and C4–2 falls between the PC-3 and Du-145

In the next step of this study, we quantitatively examined the changes in the intracellular disorder strength that may have been introduced to the three prostate cancer cell lines after prolonged exposure to docetaxel drug. We hypothesized that these cell lines might develop resistance to the docetaxel drug after 8 months of exposure and that this behavior would be associated with additional intracellular structural changes. In particular, a distinct change in the disorder strength in drug-resistant prostate cancer cells was expected. To test our hypothesis, we imaged and measured the average $L_d$ values for the docetaxel treated DU-145, C4–2 and PC-3 cells, and compared with their corresponding 8 months age-matched nondrug-treated cells (ie, drug-sensitive), obtained as described above. The bright field and PWS images (ie, $L_d$ images) of the cells are shown in Figure 4, while the comparison of their disorder strength values, $L_d$, are shown in Figure 5.

FIGURE 4.

(A-F) Bright-field images of three human prostate cell lines: control (A-F) and their corresponding drug-resistant cells from the same cell type (A’-F’). The colored images are $L_d$ images: a 2D map of intracellular $L_d$ distribution inside the normal control prostate cancer cells and the corresponding drug-resistant prostate cancer cells. The scale bar in the figures is 10 μm

FIGURE 5.

Bar plots for mean intracellular disorder strength values ($L_d$) calculated for the three prostate cancer cell lines: (A) DU-145, (B) C4–2 and (C) PC-3 (control) and the mean $L_d$ values for these cancer cells treated with docetaxel for about 8 months. Surviving cells are called drug-resistant cells, denoted as R. It can be seen that each drug-resistant cell type has higher $L_d$ than its corresponding controls. The percentage increase of disorder strengths in the drug-resistance cells, relative to their corresponding nondrug-treated DU-145, C4–2 and PC-3 cells, measured in this study are 34%, 45% and 30%, respectively. ANOVA test showed significant differences among the different groups (resistant and the nondrug-treated cells) F_{(6, 20)} = 18.953, P < 0.05). The differences between each pair group (the resistance type and their respective control) were tested using Tukey's post hoc test. Bars are means ±SE

As can be seen in Figure 5, the disorder strength of the drug-resistant cells are higher than their corresponding age matched normal (control) cancer cells. A P value <0.05 was obtained in each case, which suggests significantly elevated structural disorder in the drug-resistant cells induced with 8-month docetaxel dose, as compared with the untreated cells. This result also shows a larger spread of $L_d$ values for the drug-resisted cells compared to their untreated counter parts. As pointed out before, it was anticipated that drug-resistant cells would show measurable structural changes associated with the aggressive behavior that prostate cancer cells would have acquired after 8 months of docetaxel chemotherapy in vitro, we indeed saw an increased disorder strength in the drug-resistant cells. In addition, we were able to quantify such changes occurring in the cells in terms of the $L_d$ values, which provided a measure of degree of aggressiveness in the cells.

It should be worth pointing out here that cells behave differently in the case of drug resistance, including, for example, drug inactivation, alteration of drug targets, DNA damage repair, cell death inhibition, epithelial-mesenchymal transition and metastasis, cancer cell heterogeneity, etc. [18, 19, 26–29]. Cancer diseases comprise of a mixed genotypic cells population that respond differently to the drug treatment over chemoresistant subpopulations. Recent studies show that microfluidic approaches have been chosen to explain the complex biological mechanisms in chemoresistant cancer cells through differences in cellular stiffness after progressive drug exposure [30]. Also, seminal studies have displayed the measurement of tumorigenicity of breast cancer cells through silicon resonant microelectromechanical systems sensors and Atomic force microscopy [31]. Other physiological characteristic features that can affect the texture of the cell structure are dysregulated intracellular pH dynamics [32]. The over expression of P-glycoprotein and other ion channels in the chemoresistant cells alter the ionic efflux and influx from inside and outside of the cell [14, 33]. The altered ionic balance and variation in cell volume [34], contribute to the variation of characteristic attributes which affects the matrix stiffness and cell texture in chemoresistant cells in comparison to the control cells [35]. All these changes within chemoresistant cells enhances the randomness/disorderness, resulting in disabling of the cellular integrity [36, 37]. Furthermore, the alteration in cell volume can also affect excitability, contraction, cell proliferation, migration, receptor cycling and apoptosis. The changes in the epigenetic mechanism alter the lipid profiles that amend the biophysical properties of the drug-resistant cells [38]. Overall, as a result, these biological changes enhance the tumorigenic activities of cell through the upregulation of oncogenic genes and proteins.

5 |. CONCLUSIONS

In summary, we have applied PWS technique, to analyze structural disorder, in terms of “disorder strength ($L_d$)” of the cells, in the selected stages of prostate cancer cell lines, namely, C4–2, DU-145 and PC-3. The stages were selected because they represent known levels of tumorigenicity/aggressiveness (metastatic potential) of the prostate cancer cells, suggesting the clinical utility of measuring disorder strength of these cells. Subsequently, we also analyzed the effect of prolonged exposure of chemotherapy drug, docetaxel, on these cancer cells. In particular, we measured the changes in the disorder strength level in the cells which developed resistance toward the docetaxel drug upon an 8-month treatment (dose amount varying from 1 to 30 nM) compared to their age matched untreated cancer cells (control). The first result showed that $L_d$ values for C4–2, DU-145 and PC-3 cell lines increases with the increasing order of metastatic potential (aggressiveness/tumorigenicity) which is DU-145 < C4–2 < PC-3. This result suggested that $L_d$ is directly correlated with the metastatic potential of these prostate cancer cell lines, and it can potentially be an efficient metric to measure the aggressiveness of the prostate cancer cells. Subsequently, in the treatment with docetaxel drug, it was found that $L_d$ values (ie, disorder strength) are significantly higher for the drug-resistant cells in comparison to their untreated counterparts, indicating higher aggressiveness in the drug-resistant cells.

It should be noted that the disorder strength, $L_d$, quantifies the intracellular structural changes, by measuring refractive index fluctuations inside the cells, which in turn arises as a result of mass density fluctuation inside the cells. In that context, the present results suggest that distinct structural changes have happened in the nanoarchitecture of control and drug-resistant cancer cells and that these changes presented as mass-density fluctuation increases in surviving cells when the drug was introduced and applied for a long time. By gene mutations, surviving cells increase their aggressive behavior and, correspondingly, their biophysical property changes, leading to an increased structural disorder in the drug resistance cells. The increase in the structural disorder in drug-resistant cancer cells can be measured by $L_d$ parameter, as demonstrated in the present report, thus suggesting its utility as a potential biomarker with which to assess the efficacy of docetaxel.

These experimental results provided new insights into chemotherapy drug-resistant cells and associated increase in their structural disorder, as explained in this work. The origin of these apparent pathologically related structural changes and their correlations with the specific molecular changes, as well as their relationship to drug resistance, are important matters to be further explored in future studies.

Supplementary Material

AUTHOR BIOGRAPHIES

Please see Supporting Information online.