Telomere lengths in women treated for breast cancer show associations with chemotherapy, pain symptoms, and cognitive domain measures: a longitudinal study

Abstract

Background: Survival rates for breast cancer (BC) have improved, but quality of life post-diagnosis/treatment can be adversely affected, with survivors reporting a constellation of psychoneurological symptoms (PNS) including stress, anxiety, depression, pain, fatigue, sleep disturbance, and cognitive dysfunction.

Methods: To assess a potential relationship between telomere length (TL) and the development/persistence of PNS, we longitudinally studied 70 women (ages 23-71) with early stage BC (I-IIIA) at 5 time-points: prior to treatment (baseline), the mid-point of their chemotherapy cycle, 6 months, 1 year, and 2 years following the initiation of chemotherapy. Measures quantified included assessments of each of the PNS noted above and TL [using both a multiplex qPCR assay and a chromosome-specific fluorescence in situ hybridization (FISH) assay].

Results: Variables associated with qPCR mean TLs were age (p = 0.004) and race (T/S ratios higher in Blacks than Whites; p = 0.019). Significant differences (mostly decreases) in chromosome-specific TLs were identified for 32 of the 46 chromosomal arms at the mid-chemo time-point (p = 0.004 to 0.049). Unexpectedly, the sequential administration of doxorubicin [Adriamycin], cyclophosphamide [Cytoxan], and docetaxel [Taxotere] (TAC regimen) was consistently associated with higher TLs, when compared to TLs in women receiving a docetaxel [Taxotere], Carboplatin [Paraplatin], and trastuzumab [Herceptin] [TCH] chemotherapy regimen [association was shown with both the qPCR and FISH assays (p = 0.036)]. Of the PNS, pain was significantly negatively associated with TL (higher pain; shorter telomeres) for a subset of chromosomal arms (5q, 8p, 13p, 20p, 22p, Xp, Xq) (p = 0.014-0.047). Chromosomal TLs were also associated with 7 of the 8 cognitive domains evaluated, with the strongest relationship being noted for chromosome 17 and the visual memory domain (shorter telomeres; lower scores).

Conclusions: We showed that race and age were significantly associated with telomere length in women treated for early stage BC and that acquired telomere alterations differed based on the woman's treatment regimen. Our study also demonstrated that pain and cognitive domain measures were significantly related to telomere values in this study cohort. Expanding upon the knowledge gained from this longitudinal study could provide insight about the biological cascade of events that contribute to PNS related to BC and/or its treatment.

Keywords: Breast cancer; Chemotherapy-related cognitive dysfunction; Chromosome-specific telomere lengths; Psychoneurologic symptoms; Telomere.

Fig. 1

Fluorescence in situ hybridization (FISH) chromosome-specific telomere images. a–d The steps used for the chromosome-specific telomere assay. Individual chromosomes from each metaphase cell were evaluated based on information gained from their DAPI counterstaining appearance (a), which was used to generate a reverse DAPI banding image (b). The intensity/size of the telomere probe signal (c) was also evaluated, with a propidium iodide (PI) stain being used to delineate the chromosomal boundaries (d). e Based on the collective information gained from the reverse DAPI, FITC labeled, merged, and propidium iodide (PI) images (shown from left to right), ratio profiles were calculated for each chromosome. In image e, a representative chromosome 3 shows a short arm telomere (top of chromosome) that is larger (fluorescence intensity value of 6.63) than the long arm telomere (bottom of chromosome; value of 4.63)

Fig. 2

MMqPCR telomere data. a Average genomic telomere values at baseline were negatively correlated with age (r = − 0.37, p = 0.001 shown graphically by the trendline). Each circle (•) represents a study subject. Age is shown on the X-axis, with the baseline T/S ratio value being shown on the Y-axis. b Average genomic telomere lengths for the 72 women evaluated are shown at baseline (time-point 1) (mean = 1.542, s.e. = 0.062), mid-chemo (time-point 2) (mean = 1.534, s.e. = 0.070), time-point 3 (mean = 1.489, s.e. = 0.064), time-point 4 (mean = 1.584, s.e. = 0.068), and time-point 5 (mean = 1.583, s.e. = 0.070). The vertical bars represent standard error values associated with each mean

Fig. 3

Mean changes in chromosome-specific telomere values at baseline (T1) compared to mid-chemotherapy cycle (T2) Time-points. A negative value (shortening) was observed at the mid-chemo time-point for 38 of the 46 chromosomal arms. The X-axis shows values for each chromosomal arm. The Y-axis shows mean differences in telomere values. The bars on the histograms indicate standard errors. Abbreviations: p = short arm; q = long arm.

Fig. 4

Assessments of telomere lengths and cognitive domains. a The trajectory of average standard scores (Y-axis) over time-point (X-axis) for each of the eight cognitive domains evaluated at each of the 5 time-points in this longitudinal study included: a speed, reaction time, complex attention, cognitive flexibility, executive function, and b memory domains. c The negative log₁₀ of p values from contrast tests jointly evaluating the short arms and long arms for each chromosome. Each line represents one cognitive domain. The gray dashed line represents the corresponding value when the p value is 0.05

Fig. 5

Associations between chromosome-specific telomere length and cognitive measures. Scatter plots of visual memory scores (Y-axis) against telomere lengths for chromosome 17 (X-axis) are shown, stratified by time-point. The chromosome 17 telomere values for the short arm (open blue circle) and long arm (solid orange circle) of each individual are connected by broken lines, with the telomere values being shown on the X-axis and the visual memory scores being shown on the Y-axis. Comparing the plots from the baseline data (time-point 1) (a) to those of the mid-chemo data (time-point 2) (b), one can see the generalized pattern of telomere shortening (shift of values to left) with decreases in visual memory scores (downward shift of values with tighter clustering)