1.

Introduction

Tumor histologic grade is a recognized, but underappreciated, predictor of prognosis in breast cancer patients.^1–3 Despite the importance of grade, it has historically not been included as one of the criteria for staging of this cancer. The current TNM staging system for breast carcinoma utilizes only tumor size and extent of local invasion (T), number of axillary lymph nodes with metastases (N), and presence of distant metastases (M) as measures of the stage of the cancer.⁴ As a result, patient prognosis (and thus the need for adjuvant chemotherapy) is often assessed using only these measures. Rakha et al.¹ noted, however, that tumor histologic grade is at least as important an indicator of patient prognosis as the number of lymph node metastases and is more important than tumor size. In a review of 22,616 patients with breast cancer, Henson et al.⁵ found that patients with low-grade cancers and sizes less than 2 cm had a 99% 5-year survival despite the presence of lymph node metastases. In another study of patients with estrogen-receptor positive breast cancer and one to three positive axillary lymph nodes, the 10-year relapse rate was 43% for high-grade tumors, but only 24% and 5% for intermediate- and low-grade cancers.⁶

In 1984, Alfano et al.^7,8 showed that optical spectroscopy could be used as an alternative technique to detect cancer by determining the different emission properties of key organic biomolecules in the tissue. Fluorescence spectroscopy using excitation wavelengths from the ultraviolet light region, for example, can highlight key biomolecules such as tryptophan, reduced nicotinamide adenine dinucleotides (NADH), and collagen in the tissue. It has been reported that cancerous tissue samples have an increased amount of the amino acid tryptophan (emission intensity peak maximum at $\sim 340\mathrm{nm}$) compared with collagen or NADH (emission intensity peaks maxima at $\sim 380$ and $\sim 440\mathrm{nm}$, respectively).^9–17 Most recently, Zhang et al.¹⁷ showed, using fluorescence spectroscopy, that increased tryptophan is linked to an aggressive cancer cell line. Tryptophan is required for protein synthesis and thus increased fluorescent levels can be expected in cancerous breast tissue due to higher cell density and uncontrolled cell division. Recently, the spectral fingerprints of tissues from patients with different breast cancer histologies were examined using a novel fluorescence device known as the ${\mathrm{S}}^3$-LED ratiometer unit and an LED excitation wavelength of 280 nm.^15,16 In those reports, we showed consistently higher ratios of the 340 over 440 nm and 340 over 460 nm emission intensity peaks in malignant samples of breast cancer patients compared with their paired normal samples, regardless of the characteristics of the patient’s cancer.¹⁵

At an excitation of 280 nm, aromatic amino acids (phenylalanine, tyrosine, and tryptophan residues) can be excited; however, the emission is most likely from the indole ring of tryptophan. Tyrosine and phenylalanine will contribute significantly less to the overall emission at this excitation due to fluorescent quenching attributed to the energy transfer of phenylalanine to tyrosine and tyrosine to tryptophan. Apart from an intraprotein energy transfer from phenylalanine to tyrosine, phenylalanine is unlikely to produce fluorescence energy transfer due to a low-quantum yield. Similarly, tyrosine can be quenched due to an energy transfer from tyrosine to tryptophan residues. Therefore, tryptophan residues dominate emission and are ideal for fluorescence resonance energy transfer. At a selective longer-wavelength excitation of 300 nm, only tryptophan residues in proteins will be excited.¹⁸ From these excitation wavelengths, the longer-wavelength (red-shifted) ${{}^1\mathrm{L}}_a$ emitting state of tryptophan will occur at a maximum emission intensity peak of $\sim 340\mathrm{nm}$. Additional contributions to the emission spectra at these excitations can occur from proximal fluorophores in the tissue such as NADH (a highly fluorescent coenzyme with absorption and emission peaks at 340 and 440 to 460 nm, respectively). Table 2 highlights the absorption and emission peaks of key organic biomolecules in tissue. Due to a large spectral overlap and matching excitation–emission spectra, fluorescent resonance energy transfer from excited tryptophan (donor) to NADH (acceptor) (a dipole–dipole interaction shown by Förster to be an interaction of coupled electronic dipoles known as donor and acceptor molecules) can occur and affect the overall fluorescent spectrum of tryptophan. Studies by Torikata et al.^19,20 have shown that energy transfer from tryptophan to neighboring NADH in pig lactate dehydrogenase and malate dehydrogenase causes a decrease in its fluorescence lifetimes and leads to quenching of tryptophan fluorescence.

In this study, the spectral emission profiles from 15 patients with breast carcinoma were obtained using the conventional LS 50 Perkin-Elmer luminescence spectrometer with excitations at 280 and 300 nm. Dual excitation wavelengths were used to show energy transfers among key fluorophores. These spectral profiles were analyzed using a combination of emission intensity peak ratios. We used the ratios of fluorescence intensities at their spectral emission peaks, or spectral fingerprint peaks, at 340, 440, and 460 nm. The relative intensities or ratios ($R$) of emission intensity peaks, from these excitations, at 340, 440, and 460 nm from malignant (M) tissue samples ($R_{\mathrm{M}1}=I_{340}/I_{440}$ and $R_{\mathrm{M}2}=I_{340}/I_{460}$) and from paired normal (N) tissue samples ($R_{\mathrm{N}1}=I_{340}/I_{440}$ and $R_{\mathrm{N}2}=I_{340}/I_{460}$) were calculated. The extent of the increase in tryptophan was measured from the ratios of cancerous ratios ($R_{\mathrm{M}1}$ and $R_{\mathrm{M}2}$) over normal ratios ($R_{\mathrm{N}1}$ and $R_{\mathrm{N}2}$) to give $R_1=R_{\mathrm{M}1}/R_{\mathrm{N}1}$ and $R_2=R_{\mathrm{M}2}/R_{\mathrm{N}2}$. We report that increased tryptophan levels demonstrated by increases in $R_1$ and $R_2$ correlate strongly with high histologic grade and with large tumor size, but not with lymph node metastases, or with estrogen receptor, progesterone receptor, or HER-2-Neu receptor statuses. This suggests that measurement of tryptophan content may be useful as a part of the evaluation of patients with breast cancer.

2.

Experiment

3.

Results and Discussion

Spectral profiles from patients with breast carcinoma were acquired using the LS 50 fluorescence spectrometer at 280 and 300 nm excitations. Table 2 highlights the optimal absorption and emission peak maxima of key organic biomolecules tyrosine, tryptophan, collagen, elastin, NADH, and flavins. With light at excitation peak wavelength maxima of 280 to 300 nm, biomolecules in tissue such as tryptophan, collagen, and NADH will absorb this Gaussian-shaped excitation spectrum of light and emit a spectrum of light due to a combination of biomolecules with peak maxima at $\sim 340$, 380, and 440 to 460 nm (largely due to tryptophan, collagen, and NADH, respectively). The most intense of these peaks (due to tryptophan) is located at $\sim 340\mathrm{nm}$. A weak secondary peak at 440 to 460 nm is due to NADH and is said to occur due to resonance energy transfer of protein–NADH complexes.^18–22 A large spectral overlap and matching excitation–emission peaks at 340 nm from tryptophan and NADH, respectively, make tryptophan and NADH a good donor–acceptor pair. The Förster distance ${\overline{R}}_0$ for this pair in solution has been reported as 25 Å [where ${\overline{R}}_0$ is obtained using the average value of the orientation factor (angle between the donor and acceptor dipoles) ${\kappa }^2$, where ${\kappa }^2=\overline{{\kappa }^2}=2/3$ to account for the fast-rotatory Brownian motion which occurs during the lifetime of the donor].²³

Table 2

Absorption and emission peak maxima of key organic biomolecules.

Figure 1(a) is a microscopic image of the pathology slide ($20\times$ magnified) from malignant breast tissue from patient $n=10$. The breast cancer sample from this patient with invasive ductal carcinoma was grade 3 (high grade) and this correlates with higher ratios of emission peaks. Figures 1(b) and 1(c) show the corresponding average fluorescence spectra from normal and malignant breast tissue samples with excitations of 280 and 300 nm. A peak emission maximum can be seen at $\sim 340\mathrm{nm}$ in both the malignant and normal samples. This result is consistent with previous studies.^15,16 The emission spectra show similar peak maxima at $\sim 340\mathrm{nm}$ using an excitation of 280 and 300 nm and suggest that the fluorescence is dominated by tryptophan residues.

Fig. 1

(a) Microscopic image of the pathology slide ($20\times$ magnified) from malignant breast tissue from patient $n=10$; and (b) and (c) corresponding fluorescent spectra from malignant (MAL) and normal (NORM) breast tissues after 280 and 300 nm excitations.

Ratios ($R$) of emission intensity peak maxima from biomolecules in malignant ($R_{\mathrm{M}1}=I_{340}/I_{440}$and $R_{\mathrm{M}2}=I_{340}/I_{460}$) and normal ($R_{\mathrm{N}1}=I_{340}/I_{440}$ and $R_{\mathrm{N}2}=I_{340}/I_{460}$) samples were calculated. Figures 2(a) and 2(b) show that $R_{\mathrm{M}1}$ and $R_{\mathrm{M}2}$ from patients with high-grade breast cancer have higher bar graphs than $R_{\mathrm{N}1}$ and $R_{\mathrm{N}2}$ from excitations 280 and 300 nm. These figures demonstrate that malignant tissue samples have higher 340 over 440 nm and 340 over 460 nm ratios compared with corresponding ratios from their paired normal tissue samples.

Fig. 2

(a) and (b) Normalized ratios $R_{\mathrm{M}1}$, $R_{\mathrm{M}2}$, $R_{\mathrm{N}1}$, and $R_{\mathrm{N}2}$ from high-grade tumors with intensity peaks at excitations 280 and 300 nm, respectively, using the LS 50 spectrometer.

Normalized, averaged, and cancerous over normal ratios ($R_1=R_{\mathrm{M}1}/R_{\mathrm{N}1}$ and $R_2=R_{\mathrm{M}2}/R_{\mathrm{N}2}$) are given in Tables 3 and 4, respectively. Tables 3 and 4 compare the histologic grade with ratio results $R_1$ and $R_2$. Patients with high histologic grade breast cancer had higher $R_1$, whereas patients with intermediate-/low-grade breast cancer had lower $R_1$ results. $R_1$ for high-grade tumors, from 280 and 300 nm excitations, had average values of 2.77608 (with standard deviation of 1.36389 and $p$-value of $0.00138<0.05$) and 2.910345 (with standard deviation of 1.49592 and $p$-value of $0.00260<0.05$).

Table 3

Normalized ratios (R1) of emission intensity peaks from RM1=I340/I440 over RN1=I340/I440.

Table 4

Normalized ratios (R2) of emission intensity peaks from RM2=I340/I460 over RN2=I340/I460.

Similar results can be seen for $R_2$, with a higher overall $R_2$ for patients with high-grade cancer and a lower $R_2$ for patients with intermediate-/low-grade cancer. $R_2$ for high-grade tumors, from 280 and 300 nm excitations, had average values of 3.90183 (with standard deviation of 2.49327 and $p$-value of $0.00484<0.05$) and 3.65772 (with a standard deviation of 1.759722 and $p$-value of $0.00081<0.05$). The averages of ratios $R_1$ and $R_2$ were 3.31149 for high-grade tumors and 0.88436 for low-grade tumors. Higher ratios correlated strongly with grade 3 (high) histologic grade, while tumors that were grade 2 or 1 (intermediate/low) demonstrated low ratios.

Large tumor size also correlated with high ratios. This is not surprising as tumor size frequently correlates with the degree of tumor grade.¹ Exceptions to the association of high ratios and large tumor sizes occurred where the tumor was lower grade ($n=9$ and 15) and the histologic grade could be regarded as a confounder. Tables 5 and 6 compare the tumor size with the ratios $R_1$ and $R_2$.

Table 5

Correlation between tumor size and ratio (R1=RM1/RN1) of emission intensity peaks from RM1=I340/I440 over RN1=I340/I440.

Table 6

Correlation between tumor size and ratio (R2=RM2/RN2) of emission intensity peaks from RM2=I340/I460 over RN2=I340/I460.

With the exception of $n=9$ and 15, the average of $R_1$ and $R_2$, from 280 and 300 nm excitations at different tumor sizes, was 2.99007 with a standard deviation of 1.16402 (above 4.5 cm), 3.41884 with a standard deviation of 2.32900 (2.5 to 4.5 cm), and 0.98116 with a standard deviation of 0.20324 (0 to $<2.5\mathrm{cm}$).

Although histologic grade and tumor size (to a lesser degree) correlate with high ratios, the number of axillary lymph node metastases, a major determinant of staging, was found not to correlate with the ratios, with tumors from patients who had a small number or zero positive nodes demonstrating high ratios if high grade, and tumors from patients with a large number of positive nodes showing low ratios if they were lower grade. For example, analysis of the samples from patient $n=14$ showed high $R_1$ and $R_2$ because the cancer was high grade, despite the fact that this patient had zero positive lymph nodes. The sample from patient $n=12$ demonstrated a high ratio presumably because it was high grade, despite the fact that only one lymph node was positive for cancer. On the other hand, analysis of samples $n=4$, 7, and 11 gave low ratios presumably because their cancers were lower grade, despite the fact that each of these patients had more than 10 positive lymph nodes. No correlation of increased ratios could be found with estrogen receptor, progesterone receptor, or HER-2-Neu receptor statuses.

4.

Conclusion

A number of researchers have shown that cancerous tissues have an increase in their tryptophan content compared with normal tissues and cells.^9–17 The emission spectra show similar peak maxima at $\sim 340\mathrm{nm}$ using an excitation of 280 and 300 nm and suggest that the fluorescence is dominated by tryptophan residues. We note possible energy transfer between tryptophan (donor) and NADH (acceptor) in breast tissue samples with excitation wavelengths of 280 and 300 nm. Further studies are needed in order to determine the tryptophan residues involved in this event and the efficiency of such an energy transfer. We also report that the increase in tryptophan content is correlated with the degree of tumor histologic grade. Indeed, the difference in tryptophan content between higher and lower grade tumors that we observed in our study was marked. The averages of ratios $R_1$ and $R_2$ were 3.31149 for high-grade tumors and 0.88436 for low-grade tumors. Tumor size also correlates with increased tryptophan content, but this appears to be due to the known association of large tumor size with high histologic grade.¹ As histologic grade is becoming recognized as an increasingly important measure of prognosis for patients with breast carcinoma, the measurement of tryptophan ratios may also be useful in the assessment of these cancers. Further studies will be needed to determine whether very high breast cancer tryptophan levels in those patients with high-grade tumors might carry a worse prognosis, and thus necessitate more aggressive treatment than would be done in patients with only slightly elevated levels.