Early Detection and Diagnosis
Increased expression of the inflammatory protein YKL-40 in precancers of the breast
Article first published online: 12 JUN 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 121, Issue 7, pages 1536–1542, 1 October 2007
How to Cite
Qin, W., Zhu, W., Schlatter, L., Miick, R., Loy, T. S., Atasoy, U., Hewett, J. E. and Sauter, E. R. (2007), Increased expression of the inflammatory protein YKL-40 in precancers of the breast. Int. J. Cancer, 121: 1536–1542. doi: 10.1002/ijc.22881
- Issue published online: 24 JUL 2007
- Article first published online: 12 JUN 2007
- Manuscript Accepted: 11 APR 2007
- Manuscript Received: 18 NOV 2006
- nipple aspirate fluid;
Serum levels of YKL-40 have been associated with inflammatory diseases and breast cancer. Our purpose was to determine if YKL-40 in breast tissue, nipple aspirate fluid (NAF) and serum is (i) concentrated in NAF compared to matched serum, (ii) increased in the NAF, serum or tissue of women with biopsy proven precancer or cancer compared to healthy women and (iii) influenced by menopausal status. 118 women (61 healthy subjects, 10 with precancer and 47 with breast cancer) aged 17–95 years provided NAF with or without serum samples for analysis. Matched tissue was analyzed from a subset of subjects who underwent breast biopsy. All NAF and serum samples had detectable levels of YKL-40. Median YKL-40 levels for the entire cohort were 683 fold higher in NAF than serum. Premenopausal subjects had higher NAF and lower serum levels of YKL-40 than postmenopausal subjects. YKL-40 levels in NAF but not serum were higher in women with precancer (atypical hyperplasia and lobular carcinoma in situ) than in either healthy subjects (p = 0.025) or subjects with breast cancer (p = 0.015). In women with precancer, YKL-40 distribution in tissue correlated with YKL serum level (p = 0.043). YKL-40 is concentrated in NAF, with the highest concentrations in premenopausal women. NAF levels of YKL-40 are significantly higher in women with precancers than healthy subjects, suggesting that measuring YKL-40 in NAF may improve the identification of women at increased breast cancer risk. © 2007 Wiley-Liss, Inc.
There is increasing evidence that inflammation is linked to the development of cancer.1 One of the best examples of this phenomenon is the inflammation that occurs with chronic hepatitis B infection, which frequently leads to hepatocellular cancer. Other examples include the ulcer causing bacterium Helicobacter pylori and the immune disorder ulcerative colitis, which increase subjects' risk of gastric and colon cancer, respectively.1 Researchers estimate that inflammation contributes to the development of at least 15% of all cancers. It appears that inflammation associated proteins are involved,1 probably through the secretion by inflammatory cells of these proteins near the tumor, and/or by increased production and secretion of the proteins by tumor cells, leading to increased tumor growth.
YKL-40 (GenBank accession no. M80927, listed as human cartilage glycoprotein-39) is a secretory molecule related in amino acid sequence to the chitinase family of proteins, but without chitinase activity.2 A 39 kDa protein was detected in the mammary secretions of nonlactating cows, but not in cow milk.3 This protein,4 as well as a 40 kDa protein labeled mammary gland protein (MGP), which is the homologue of YKL-40 in the goat, were subsequently found to increase during the time of breast involution, after nursing ends and the breast returns to the prepregnancy state.5 MGP-40 has significant homology with YKL-40,5 as well as mammalian chitinase-like proteins,6 which include YKL-39. A murine homologue of YKL-40 in mice known as breast regression protein (BRP)-39 is induced in mammary epithelial cells a few days after weaning. Murine mammary tumors initiated by neu or ras oncogenes were found to express BRP-39, although tumors initiated by c-myc or int-2 oncogenes did not.7 It has been proposed that MGP-40 normally acts as a protective signaling factor that determines which cells survive during breast involution, but that cancers might use this protein, or related proteins, to extend cell survival.5
Serum YKL-40 increases linearly in healthy subjects up to age 79.8 Preliminary investigations of YKL-40 demonstrated its role in rheumatoid and osteoarthritis. The protein is produced by human chondrocytes and synoviocytes, and the levels of the protein before treatment of rheumatoid arthritis (RA) are significantly higher than healthy controls.9 YKL-40 levels are 10-fold higher in synovial fluid than in serum.8 Serum YKL-40 levels correlate with C-reactive protein and the number of swollen joints in RA.9
It has become increasingly clear that circulating levels of YKL-40 are related to poor survival in subjects with breast,2, 10, 11 colorectal.12 glioblastoma,13, 14, 15 lung16 and ovarian17 cancer, as well as melanoma.18, 19 YKL-40 is expressed not only by inflammatory cells, but also by clinical cancer tissues.20, 21
Currently available breast cancer screening tools such as mammography and breast examination miss 10–40% of early breast cancers. An invasive needle or surgical biopsy must be performed when an area of suspicion is identified in order to confirm, by cytologic or histologic evaluation, the presence of malignancy, even though 66–85% of abnormalities are benign.22 The development of noninvasive techniques that would distinguish between women with or without breast cancer is therefore of crucial importance.
Nipple aspiration is a noninvasive, low cost procedure that provides a relatively small set of breast specific proteins. We are able to obtain NAF in 97% of pre- and postmenopausal subjects.23 Because the proteins are secreted, they represent the final processed form of the protein, which makes protein analyses less ambiguous. It can be collected noninvasively with minimal or no discomfort,23 making it potentially ideal to evaluate the effect of a chemopreventive or a chemotherapeutic agent. Validation of nipple aspiration as a technique to screen for breast cancer requires the demonstration that one or more biologic markers in the fluid, such as YKL-40, correlate with breast tissue pathology and thus predict disease.
Previous studies addressed the association of serum YKL-40 in breast cancer patients with metastases to nodes, bone or viscera2, 10 and the association of YKL-40 levels with disease free survival.11 The focus of the current study was to analyze both systemic (serum) and an organ specific (NAF) body fluids and breast tissue, to determine if YKL-40 levels in the specimens were (i) concentrated in NAF compared to matched serum, (ii) increased in the progression of disease risk from healthy women, women with precancer, to women with breast cancer and (iii) influenced by menopausal status.
Material and methods
Women were required to give written informed consent to an Institutional Review Board approved protocol in order to enroll in the study. Pregnant and lactating women were not eligible. Information on arthritis and NSAID use was obtained from the participant's self report or a physician note. On the basis of whether the subject had biopsy proven newly diagnosed breast cancer, biopsy proven evidence of a diagnosis which increased their risk of breast cancer, or neither of these, enrolled subjects were classified into 1 of 3 risk categories: cancer, if there was evidence of ductal carcinoma in situ (DCIS) or invasive breast cancer (IBC); precancer, if the subject had atypical duct hyperplasia (ADH) or lobular carcinoma in situ (LCIS); or benign, if neither precancer nor cancer was present. NAF and tissue were analyzed from only one breast. For subjects with precancer or cancer, NAF and tissue were analyzed from the breast with known disease.
NAF samples using a modified breast pump23 were collected into capillary tubes and stored at –80°C until use. Eight milliliter of blood were also collected from the subject in a red top tube, the blood spun for 10 min at 1,600 rpm, the serum fraction decanted and stored at –80°C until use. Paraffin embedded, formalin fixed tissue was collected from the pathology department.
NAF and serum
These samples were analyzed by immunoassay for their YKL-40 content as per the manufacturer's instructions (Quidel Corporation, San Diego, CA). The kit uses a monoclonal anti-YKL-40 antibody conjugated to biotin that binds to streptavidin and captures YKL-40 in the standard or sample. Briefly, 20 μl of a diluted NAF or serum sample was pipetted into an appropriate well, 100 μl of capture solution added, the mixture incubated for 1 hr at room temperature, washed, substrate solution added followed by 1 hr incubation and absorbance measured at 405 nm.
A standard curve was prepared using serial dilutions of YKL-40. A linear regression equation was created from standards of known YKL-40 concentration, and YKL-40 concentrations of unknown samples fit to the standard curve regression equation, corrected for aliquot volume and expressed as nanograms of YKL-40/ml of original sample. Specifically, the dilutions were 300, 200, 100, 50, 20, 0 ng/ml. Relative absorbance vs. known YKL-40 concentration was plotted. Once the standard curve was produced, the relative absorbances of the unknown samples were compared to the standard curve and the number of nanograms of YKL-40 in each unknown tube calculated. The standard curve, NAF and serum samples were run in duplicate and the average of the 2 values reported. The goodness of fit, R2, for both NAF and serum samples was 0.995. Total protein concentrations in NAF samples were determined using the bicinchoninic protein assay kit from Pierce (Rockford, IL). NAF results are presented based on total protein concentration. The ratio of NAF/serum YKL-40 compares NAF concentration in ng/ml to serum values in ng/ml.
After routine processing, the formalin fixed, paraffin-embedded breast tissue was cut at 3 μ thicknesses onto glass slides and heated for 90 min in a 60°C oven. The slides were deparaffinized and hydrated with water. The sections were then pretreated with steam/EDTA for 20 min followed by room temperature for 20 min. A 3.8 mg/ml YLK-40 mAb antibody 207F9 (kind gift of Dr. Paul Price, University of California-San Diego) which has been previously shown to work in human breast tissue6 was diluted 1:60. The primary antibody was applied for 30 min followed by a secondary antimouse antibody for 30 min. The DAKO Envision Plus automated staining system was used (DAKOCytomation, Carpenteria, CA). Following staining, the slides were counterstained with Harris Hematoxylin for 1 min and coverslipped. The negative control substituted murine IgG (Vector Laboratories, Burlingame, CA) for the YKL-40 mAb. The positive control was a tissue section containing inflammatory infiltrates and cutaneous sweat glands known to react with YKL-40. Immunohistochemical grading was performed by Dr. Timothy Loy, director of Anatomic Pathology at the University of Missouri School of Medicine, and Ronald Miick, senior resident in Pathology, using 2 parameters, intensity (0, no staining; 1, weak; 2, moderate; 3, strong) and distribution (0, no stain of lesional cells; 1, up to 33% of lesional cells stained; 2–34, 66% of lesional cells stained; 3–67, 100% of lesional cells stained).
For YKL-40 levels in NAF and serum, median values of continuous variables were computed for the various groups of subjects. Because of the potential nonnormality of the data, ranking procedures were used for all analyses with continuous variables. The Wilcoxon Rank Sum Test was used to compare independent groups. Examples of these comparisons include comparing pre- and postmenopausal women, comparing the cancer to the precancer group, etc. The Wilcoxon Signed Ranks Test was used for within group comparisons.
For the assessment of YKL-40 expression in tissue, subjects were grouped according to cancer risk level. The objective was to determine if the levels of YKL-40 intensity and distribution differed relative to cancer risk. When 2 groups were compared the Wilcoxon Rank Sum test was used since the 2 variables were measured at the ordinal level. When 3 or more groups were compared the Kruskal-Wallis Test was used. Spearman correlations were computed to determine if distribution and intensity were correlated and if these 2 variables were correlated with YKL-40 in NAF and serum.
One hundred eighteen women aged 17–95 were enrolled in an Institutional Review Board approved protocol to evaluate the expression of YKL-40 in NAF and serum samples (Table I). Both NAF and serum were analyzed for YKL-40 when available. Median age of the subjects ranged from 46 years in the benign to 58 years in the cancer groups. In both groups most subjects were postmenopausal and Caucasian. Because of their known or suspected influence on YKL-40 expression, the following clinical variables were analyzed: menopausal status, presence or absence of arthritis, and use of nonsteroidal anti-inflammatory medications.
|Premenopausal||28 (45.9)||4 (40)||14 (29.8)|
|White||57 (93.4)||10 (100)||44 (93.6)|
|Black||3 (4.9)||0||3 (6.3)|
|Documented arthritis1||13 (21)||4 (40)||8 (17)|
|NSAIDs2||22 (36.1)||4 (40)||15 (31.9)|
YKL-40 levels are concentrated in NAF
As already mentioned, YKL-40 is concentrated in synovial fluid compared to matched serum in subjects with arthritis.8 In this study, we determined if YKL-40 was concentrated in the breast by comparing levels in NAF to those in matched serum (Table II). Whereas the NAF results are expressed in ng/mg, controlling for total protein concentration which can vary from sample to sample, the NAF/serum YKL-40 assessment used levels in ng/mL for both body fluids, so that it was an appropriate comparison. Overall, YKL-40 was concentrated in NAF 683 fold. The ratio of NAF/serum YKL-40 was highest in women with precancer (1,046) and lowest in women with cancer (156) (Table II, Fig. 1). These concentration differences based on disease risk were not significant.
|N (ng/mg)||N (ng/mL)||N||Ratio|
|Overall||118 (278)3||83 (55)||83||683|
|Benign||61 (275)||45 (45)||45||829|
|Precancer||10 (602)||9 (60)||9||1,046|
|Cancer||47 (186)||29 (67)||29||156|
|Noncancer4||71 (296)||54 (46)||54||853|
|Cancer||47 (186)||29 (67)||29||156|
|Benign||61 (289)||45 (45)||45||829|
|Precancer||10 (602)||9 (60)||9||1,046|
|Precancer||10 (602)||9 (60)||9||1,046|
|Cancer||47 (186)||29 (67)||29||156|
|Local disease only|
|HER2+||3 (3.3)||2 (151)||2||1.7|
|HER2−||8 (321)||7 (26)||7||853|
|HER2+||2 (529)||1 (124)||1||43|
|HER2−||10 (533)||8 (62)||8||573|
|HER2−||1 (416)||1 (129)||1||137|
Of 47 patients with breast cancer, 28 had tumor confined to the breast, 18 had spread to axillary lymph nodes and one had distant disease spread. Heregulin (HER)2 results were available for 24 patient specimens with breast cancer. HER2 expression is not evaluated at our institution on normal or precancerous specimens, and was not available. Of the 24 cancers with HER2 results, 5 were HER2 positive and 19 were HER2 negative. There was a trend for HER2 positive tumors to have lower NAF and serum values of YKL-40 (Table II).
The ratio of NAF/serum YKL-40 was greater (Table III) in pre- than in postmenopausal women (p = 0.0014). This difference was most pronounced in women with cancer, where the ratio of NAF/serum YKL-40 was 1652 in premenopausal women vs. 137 in postmenopausal women (p = 0.0086).
|Menopausal status||Pre||(N)||Post||(N)||p value|
Breast cancer risk influences YKL-40 expression in NAF but not serum
We divided subjects into 3 categories, benign, precancer and cancer, to assess if NAF YKL-40 was associated with cancer risk. There was a significant difference between the groups (p = 0.027). NAF levels were highest in breasts with precancer, higher than in benign breasts (p = 0.025) and higher than in breasts with cancer (p = 0.015) (Table II, Fig. 1). Levels in DCIS and IBC were not significantly different.
YKL-40 expression in tissue
Twenty-eight subject specimens were analyzed for YKL-40 expression in tissue (Table IV, Fig. 2). These specimens were randomly chosen from enrolled subjects who had undergone breast surgery after NAF and serum collection. YKL-40 was expressed in 3/3 benign, 8/8 precancerous and 16/17 cancerous breast tissue specimens. The staining pattern was predominantly cytoplasmic, particularly in cases with less than 100% staining distribution patterns. Some cases demonstrated nuclear as well as cytoplasmic staining. Blood vessels and inflammatory cells also had YKL-40 immunoreactivity. Staining intensity and distribution were correlated (p < 0.01). Neither staining intensity nor distribution was influenced by disease diagnosis. In tissues containing a precancerous lesion but not in other tissues, YKL-40 levels in serum correlated directly with YKL-40 tissue distribution (N = 8, R = 0.72, p = 0.043).
|Hyperplasia||3||2.33 ± 0.58||1.67 ± 0.58|
|Atypical hyperplasia||7||2.57 ± 0.53||2.14 ± 0.69|
|Lobular carcinoma in situ||1||2||1|
|Ductal carcinoma in situ||5||2.60 ± 0.89||1.60 ± 0.55|
|Invasive breast cancer||12||2.00 ± 1.04||1.83 ± 0.72|
We have estrogen and progesterone receptor (ER/PR) as well as HER2 expression results for all 12 invasive breast cancers (Table V). We also have ER/PR information for 2 of 5 cases of DCIS and HER2 information for 1 of 5 cases of DCIS. There was neither association of YKL-40 staining distribution nor staining intensity with ER/PR or HER2 expression.
YKL-40 levels in NAF and serum are influenced by menopausal status
Levels of YKL-40 in NAF were lower (p = 0.0035) and in serum were higher (p = 0.012) in post- than in premenopausal women (Table III). Although the trends of all subgroups (benign, precancer, and cancer) paralleled the overall findings of lower values in NAF and higher values in serum in post- than in premenopausal women, none of the subgroup differences were significantly different.
Age influences YKL-40 in NAF and serum differently
Because of the association of YKL-40 levels with inflammation due to arthritis and other diseases of aging, we assessed whether age was associated with YKL-40. Serum levels of YKL-40 were associated with age in postmenopausal women (R = 0.50, p = 0.0001), but not in premenopausal women. NAF levels of YKL-40 were not associated with age in either pre- or postmenopausal women. We also assessed if women with documented arthritis had different levels of YKL-40 than those without, and found that neither NAF nor serum levels were significantly different.
Most reports of serum YKL-40 and its association with cancer have focused on later stage disease and prognosis. We are aware of 3 reports that evaluated patients with early stage disease. Higher levels of YKL-40 were found in patients with ovarian cancer than in healthy women or women at high risk for the disease.17 High levels of the protein were also found in women with breast cancer who had a shorter recurrence free survival11 and in patients with early stage melanoma who had shorter relapse-free and overall survival.19 We determined if YKL-40 levels in serum, a systemic fluid contributed to by many organs, as well as two breast specific specimens, NAF and tissue, would predict if a woman had precancer or cancer.
First, we assessed if YKL-40 was concentrated in NAF relative to serum, as had been reported in synovial fluid.8 We report, to our knowledge for the first time, that YKL-40 is present and highly concentrated in NAF, but unlike in synovial fluid,24 levels in NAF and serum appear to be inversely related, with higher levels in NAF in premenopausal women, but higher levels in serum in postmenopausal women. In premenopausal women, NAF levels of YKL-40 trended higher than in serum in all disease groups, benign, precancer and cancer suggesting that YKL-40 in NAF is derived, at least in part, from noninflammatory breast cells. YKL-40 might promote tumor initiation in precancers through the secretion by inflammatory cells and/or by increasing the production and secretion of tumor promoting proteins. YKL-40 may also function, along with other growth factors, to aid precancerous cells in their growth and metastasis. This would explain why it is elevated in women with precancer, as this is a period in which the transformed cells can either progress to fully malignant cells, regress toward a benign phenotype, or undergo apoptosis.
The link between inflammation and cancer is an area of active research. It appears that molecules such NF-κB, which are involved in orchestrating the immune response, may under some circumstances actually promote cancer cell growth and metastasis by inhibiting malignant cell apoptosis.25 Though the immune system often plays an active anticancer role, it may under certain conditions promote cancer growth. Macrophages, a key component of the immune system, secrete a variety of growth factors when summoned by the immune system into an area of tissue remodeling. In a milieu bathed in a variety of growth factors, precancerous cells may take advantage of this situation to mutate and grow.
We were surprised that NAF levels of YKL-40 were higher in precancers than either levels from healthy women or women with cancer, given that YKL-40 circulating levels have been reported to be higher in women with cancer than healthy women, and higher in LN positive than LN negative patients.11 Perhaps this is not as surprising as expected, given that YKL-40 levels rise after pregnancy during breast involution,11 a process whereby many of the ductal epithelial cells which proliferate during pregnancy and lactation undergo apoptosis. It is possible if YKL-40 expression in the breast assists in the involution process, perhaps by acting as a protective signaling factor that determines that cells survive during breast involution, and may thereby influence future breast cancer risk. Additionally, YKL-40 production by normal breast epithelial cells in healthy premenopausal women may play a role in the normal turnover of these cells during the 28 day hormonal cycle.
Prior reports indicated that the highest levels of serum YKL-40 among patients with breast cancer are in women with distant disease spread,2 followed next by levels in women with disease spread to axillary lymph nodes, then women without spread beyond the breast.11 We observed a similar trend, with median serum YKL-40 values increasing from 42.9 ng/ml for women with local disease to 73.3 ng/ml for women with nodal spread, and the single woman with distant metastases having a value of 129.4 ng/ml. Although we did not observe a significant association between serum YKL-40 and breast cancer, median levels did trend higher with disease progression from benign to cancer (Table II).
YKL-40 tissue expression has been evaluated in a variety of human tumor types, including brain and small cell lung. YKL-40 immunostaining appears to be more common (95% vs. 38%), and more likely to have moderate-intense staining (84% vs. 24%), in glioblastomas than in anaplastic oligodendrogliomas.20 Protein expression was also evaluated in small cell lung cancer specimens, with no tumor cells expressing mRNA, whereas mRNA was detected in the peritumoral stroma in 90% of specimens.26 We observed that YKL-40 was expressed in all benign and precancerous, and in 88% of cancerous breast tissue specimens. This is in contrast to the brain tissue studies, in which normal brain did not stain for YKL-40. On the basis of observations in brain, lung, and breast cancer, it appears that YKL-40 expression in malignancies and in benign specimens is tissue specific. We also observed YKL-40 expression in precancerous tissue which trended higher than in normal or cancerous tissue. Whether this is real and limited by sample size is unclear. Our observation of both nuclear and cytoplasmic YKL-40 staining in a minority of cases differs from previous immunohistochemical studies, which observed only cytoplasmic staining. The reason for the nuclear staining seen in some of our cases remains unclear. The predominant cytoplasmic staining which we observed suggests that the protein of YKL-40 concentration is greater in the cytoplasm than in the nucleus. It is possible that better antigen preservation in our cases, or a more sensitive immunohistochemical technique allowed us to observe nuclear staining for YKL-40 which was not previously described. The interesting observation that the fraction of precancerous cells staining for YKL-40 correlates with serum levels of the protein requires validation before drawing firm conclusions.
Although we did not find an obvious association in tissue between YKL-40 and ER, PR or HER2 expression, there was a trend for HER2 positive tumors to have lower YKL-40 NAF values and higher serum values, especially among women with disease confined to the breast. This is consistent with the observation that murine mammary tumors initiated by neu or ras express the murine YKL-40 homologue BRP-39, while tumors initiated by c-myc or int-2 do not.7
Neither the presence of arthritis nor NSAID use influenced YKL-40 levels in either NAF or serum. On the other hand, YKL-40 levels in the serum of post- but not premenopausal women correlated directly with age. Thus, at least in our participants, in postmenopausal women age was a better predictor of YKL-40 circulating levels than was information on arthritis or NSAID use. It is also possible that YKL-40 metabolism on may not be optimally inhibited by NSAIDs, or different classes of NSAIDs may have different effects upon YKL-40 metabolism. This is an area which requires further study.
In summary, YKL-40 is secreted by normal cells responding appropriately to stimuli which trigger tissue remodeling, mediated by the immune system. Unfortunately, this may unwittingly also promote cellular transformation from a precancerous to a cancerous state. A high NAF level of YKL-40 in an apparently healthy woman, or levels which rise over time, may prove useful in risk stratification.
We thank Dr. Paul Price for providing the mAb for tissue assessment of YKL-40 and Dr. Julia Johansen for providing information on YKL-40 physiology. We also thank histotechnologists Ms. Carolynn Mision and Ms. Nancy Dawson for optimization of the YKL-40 immunostain of the formalin- fixed breast tissue.
- 21YKL-40 expression in benign and malignant lesions of the breast. Proc ASCO 2005; 9665A..