Abstract
Background/Aim: Regarding the impact of tumor-infiltrating immune cells on tumor cells, many contradictory reports have been published. We have hypothesized that these controversies result from differences in tissue types and tumor stages, in which immune cells are variably distributed and differentially associated with epithelial cells. Our current study compared the pattern and frequency of physical association of tumor-infiltrating immune cells with different parenchymal cells of human breast and prostate tumors harboring normal, hyperplastic, in situ, and invasive components. Materials and Methods: The cytological, biological, and molecular alterations were assessed with double immunohistochemistry, double fluorescent labeling, apoptosis assay, and gene expression profiling. Results: Our study detected several previously undescribed features: (i) over 95% of infiltrating immune cells were seen within normal, hyperplastic, or in situ cancer structures with focally-disrupted capsules, and fewer than 5% were found within invasive cancer; (ii) over 95% of normal, hyperplastic, and in situ cancerous epithelial cells were physically shielded from immune cells by the surrounding myoepithelial or basal cell layer; (iii) about 90% of myoepithelial or basal cells physically associated with immune cells and such residual cells within focally disrupted layers exhibited distinct degeneration, including apoptosis, necrosis, and reduced expression of tumor suppressor p63; (iv) epithelial cells overlying focally disrupted tumor capsules surrounded by immune cells had substantially higher proliferation than their adjacent counterparts, and some of the proliferating cells were arranged as tongue-like projections invading the stroma; and (v) microdissected cells overlying focally disrupted tumor capsules had more than 5-fold higher expression of stem cell lineage markers KIT and NCOR2. Conclusion: Tumor-infiltrating immune cells are primarily associated with degenerated myoepithelial or basal cells causing focal disruptions of the capsule, which selectively favor proliferation, invasion, and dissemination of the overlying tumor stem cells.
Many contradictory reports have been published regarding the impact of tumor-infiltrating immune cells on tumors. On one hand, certain studies have reported that direct physical contact between infiltrating immune cells and tumor cells is associated with the destruction of associated tumor cells, reduction of tumor size, and significantly improved clinical prognosis (1-6). On the other hand, a steadily increasing number of publications have shown that increased immune cell infiltration promotes tumor progression and invasion (7-15). It has been reported that stage- and histopathologically-matched pre-invasive prostate and esophageal tumors with increased immune cell infiltration have a significantly higher frequency of subsequent progression to invasive lesions than their counterparts without immune cell infiltration (7-9). It has also been suggested that aberrant immune cell infiltration promotes tumor metastasis through a number of mechanisms: macrophages can enhance cancer cell migration through secretion of chemotactic and chemokinetic factors that promote fibrillogenesis and angiogenesis, allowing tumor cells to track along collagen fibers to blood vessels (10, 11); macrophages ingest tumor cells, resulting in the mixture of genetic material that creates a hybrid phenotype, which allows metastasis to form new colonies to remote sites (12); T-lymphocytes promote invasion and metastasis by regulating tumor-associated macrophages (13); and infiltrating immune cells can export growth factors and other proliferation-related molecules to associated tumor cells through direct physical contact, promoting epithelial–mesenchymal transition and cell motility (14, 15).
The primary reason for these contradictory observations and the molecular mechanism(s) for the reported diverse impacts, however, have not been elucidated. In addition, immune cell-mediated cytotoxic assays during the past were predominantly, if not solely, conducted on cell cultures or animal models without the inclusion of human tumor tissues for comparison (16-18). As the immune surveillance systems in human and animal tissues differ significantly (19, 20), it is not known if these in vitro assay's results truly reflect the intrinsic impact of tumor-infiltrating immune cells on tumors.
It has long been our speculation that these contradictory reports may result from differences in tissue types and tumor stages, in which infiltrating immune cells are differently distributed and physically associated with different cell types. The normal and pre-invasive breast or prostate epithelium is surrounded by the basement membrane and a single layer of elongated cells, called basal cells in the prostate and myoepithelial cells in the breast. The basal or myoepithelial cell layer covers the epithelial cells on one side and rests on the basement membrane on the other. The basement membrane and the basal or myoepithelial cell layer collectively constitute a continuous capsule that segregates epithelial cells from stromal and vascular structures (21-24). The epithelium is normally devoid of blood vessels, lymphatic ducts, and infiltrating immune cells. The secretory products of normal epithelial cells and the debris of aged and injured epithelial cells directly drain to the acinar or ductal lumen. In contrast, the secretory products of normal basal or myoepithelial cells and the debris of aged or injured myoepithelial cells or basal cells may easily diffuse to the adjacent stromal and vascular structures (25), which can stimulate the production of auto-antibodies or attract infiltration of immune cells to the affected site.
Based on results from our recent studies on multiple types of human cancer, including breast, prostate, cervical, lung, skin, colorectal, and oral cancer, we have derived a novel hypothesis that tumor invasion and metastasis arise from the convergence of a tissue injury, the innate immune response to that injury, and the presence of tumor stem cells within the tumor capsules at the site of the injury (26-37). According to our hypothesis, focal capsule degeneration due to age or disease attracts infiltration of lymphocytes, which digest and remove degenerating capsules, causing the formation of focal disruptions in the capsule. If the focal disruption occurs in the capsule underlying progenitor or tumor stem cells, these cells are afforded the opportunity to proliferate, leading to their ‘budding’ and dissemination from the focal disruption, and subsequent growth into the surrounding stroma and vascular structures (26-37). Our hypothesis has been recognized as being more compatible with existing experimental evidence than the traditional ‘protoelytic enzyme’ theory for tumor progression and invasion (38-40).
Our current study attempted to further validate our hypothesis. Since cytotoxic lymphocytes and natural killer cells are the primary infiltrating immune cell types, which have to physically contact their targets in order to exert their cytotoxic functions (41-44), our current study anticipated that tumor-infiltrating immune cells would be exclusively or preferentially distributed within the tissue components with focally disrupted tumor capsules. Consequently, cytological signs of degeneration would be exclusively or preferentially seen in basal or myoepithelial cells physically associated with infiltrated immune cells, whereas invasion and metastasis-related changes would be exclusively or preferentially seen in luminal cells associated with degenerated myoepithelial or basal cells.
Materials and Methods
Unpublished experimental data, hematoxylin and eosin (H&E), and immunohistochemical photographs from our previous studies (26-37) were re-examined. These studies included over 1,000 cases of human breast and about 100 cases of prostate tumors predominantly harboring pre-invasive lesions with a minimal invasive tissue component (accounting for no more than 10% of the entire tumor cell population). These studies were primarily aimed at the identification of the early morphological and immunohistochemical signs and consequences of tumor capsule disruptions and lymphocyte infiltration. The original tissue blocks and unstained tissue sections were obtained from the archives of the Armed Forces Institute of Pathology (which was disestablished in accordance with BRAC Legislation on September, 2011), and from our collaborative projects with the National Cancer Institute, Thomas Jefferson University Medical School, Howard University Hospital, Georgetown University Lombardi Cancer Center, Fairfax Hospital, and Beijing 301 Hospital. All cases were sent to Armed Forces Institute of Pathology and collaborators for initial or conformational diagnosis, and none was known to have received any prior chemotherapy. These studies were conducted with multiple IRB (UBNI, UBJX, UBZP2, 05AA, and 07DJ from the Armed Forces Institute of Pathology) approval and were financially supported by a number of governmental and private organizations (see Acknowledgements). The results have been approved for publications by these funding organizations and other authorities.
Serial 5-7 μm sections were made from formalin-fixed, paraffin-embedded breast and prostate tumor tissue blocks, and the first and last sections from each block were stained with H&E for morphological classification using published criteria (26). Double-immunohistochemistry was used to assess the potential impact of infiltrating immune cells on other cell types using our published protocol (45). The secondary antibody, ABC detection and diaminobenzidine chromogen kits were obtained from Vector Laboratories (Burlingame, CA, USA). The AP red-chromogen kit was purchased from Zymed Laboratories (South San Francisco, CA, USA). To assess the specificity of the immunostaining, different negative controls were used, including the substitution of the primary antibody with the same isotype or pre-immune serum of the antibody; and omission of the secondary antibody. The immunostaining procedure was repeated at least twice using the same protocol and under the same conditions.
Immune cell aggregates and tumor-infiltrating immune cells were first elucidated with a mouse monoclonal antibody to human leukocyte common antigen (LCA; clone: 2B11+PD7/26; Dako, Carpinteria, CA, USA), which is expressed in all normal hematopoietic cells and their neoplastic transformations. To identify the specific types of infiltrating immune cells, a variety of antibodies to markers, including CD3, CD4, CD8, CD 16, CD20, CD56, FOX3P, mast cells and neutrophils were used. Immune cell aggregates were defined as lymphoid masses with at least 100 individual lymphocytes/per aggregate physically associated with or immediate subjacent to each other in a confined field.
The human breast and prostate epithelial cells were identified with a mouse monoclonal antibody to human cytokeratin (CK) AE1/AE3 (clone: AE1/AE3; DakoUSA), which reacts with all epithelium-derived cells. The human breast myoepithelial and prostate basal cells were identified with mouse monoclonal antibodies against human smooth muscle actin (SMA; clone: 1A4; Sigma, St. Louis, MO, USA) and CK 34βE12 (clone: M0630; Dako), respectively. Focal capsule disruptions were defined as the presence of a physical gap in a given myoepithelial or basal cell layer larger than the combined size of at least three epithelial cells in at least three consecutive sections. Physical signs of degeneration-related changes included the loss of expression of phenotypic markers, vacuolation, fragmentations, swelling, nuclear membrane breakage, chromatin condensation, atrophy, and necrosis.
To compare the frequencies of cytological signs of degeneration in basal (or myoepithelial) and luminal cells physically associated with infiltrated immune cells, digital images were taken from each case in 3-5 randomly selected basal (or myoepithelial) cell layers and associated luminal cells immediately subjacent to or surrounded by, infiltrated immune cells. Digital images were enlarged and reviewed on a computer screen to detect potential signs of degeneration defined above by three investigators (Drs. Man, Song, and Chen). Only totally agreed data by three investigators were included. A given cell was considered immunoreactive if distinct immunoreactivity was consistently seen in its cytoplasm, membrane, or nucleus, whereas all negative controls lacked distinct immunostaining. All infiltrated immune cell aggregates in each case were counted, and the frequencies of these aggregates within pre-invasive and invasive tissue components were statistically compared with the Pearson's Chi-square test. The total numbers of cells encountered and cells with cytological signs of degeneration from each case of the same category were added and averaged, and the averaged frequencies were statistically compared with the Pearson's Chi-square test. Statistical significance was defined as p<0.05.
To further identify the biological and molecular consequences along with the potential mechanistic mechanisms resulting from the physical association between infiltrating immune cells and the luminal or basal (myoepithelial) cells, the following technical approaches or assays were utilized.
Apoptosis assay. Tissue sections from 10 cases with a high frequency of degeneration-related changes in the basal cell layers were subjected to apoptosis detection with an ApopTag® Plus Peroxidase In Situ Apoptosis Detection Kit from CHEMICON International (Temecula, CA, USA) according to the protocol provided by the manufacturer. After the apoptosis assay, the sections were thoroughly washed with PBS and subjected to immunohistochemical staining for CK 34βE12 to determine the histological origin of apoptotic cells.
The expression status of tumor suppressor p63 in basal cell layers. The expression status of p63 in basal (or myoepithelial) cells within non-disrupted and focally disrupted residual layers were compared in sections double-immunostained for CK 34βE12 and p63. To verify the subcellular localization of basal cell phenotypic markers CK 34βE12 and p63, sets of adjacent human prostate tissue sections from 10 selected cases were double-immunostained for CK 34βE12 and p63 (clone: 4A4; Cell Marque, Rocklin, CA, USA). The antigen–antibody complexes were distinguished with different chromogens and secondary antibodies labeled with different fluorophores (DyLight 488 and Dylight 549; KPL, Gaithersburg, MD, USA), respectively, according to the manufacturer's instructions. Immunostained sections were examined and digital images were taken in a fluorescent microscope (Fluoview 300; Olympus America, Inc., Center Valley, PA, USA).
Proliferation index and lymphatic duct density. To further assess the potential impact of infiltrating immune cells on cell proliferation and angiogenesis, tissue sections from 30 selected cases were double immunostained for LCA and a cell proliferation marker, Ki-67 (clone: MM1; Dako,) and a lymphatic endothelial marker, D2-40 (clone: D2-40; Signet, Dedham, MA, USA). The proliferation status and the lymphatic duct density in epithelial cells surrounded by and distant from infiltrating immune cells were compared. In addition, special attention was paid as to whether proliferating cells were preferentially located at the site of focally disrupted epithelial capsules.
Gene expression of vrofilin2. Similar immunohistochemical methods as those described above were applied to freshly-frozen human breast (N=20) and prostate (N=20) tissues harboring mainly pre-invasive lesions with a minimal invasive tissue component for gene expression profiling. From each case, 20 consecutive sections at 10-12 μm thickness were prepared, and sequentially placed on positively charged microscopic slides. Sections 1, 10, and 20 of each case were double immunostained, which allows simultaneously elucidation of the tumor cells and the tumor capsules as described above. The remaining sections were lightly stained with hematoxylin and used as references. Tumor cell clusters overlying focally disrupted capsules and their adjacent counterparts within the same tumor core at a distance from the disruptions were microdissected, and subjected to RNA expression with a PicoPure™ RNA isolation kit from Arcturus Bioscience, Inc. (Mountain View, CA, USA) using the protocol provided by the manufacturer. Extracted total RNAs were linearly amplified using RiboAmp® OA 1 Round RNA amplification kits (Arcturus Bioscience, Inc.). Amplified antisense RNA (aRNA) products were converted to biotin-labeled cDNA with an AmpoLabeling kit (SuperArray Bioscience Corporation, Frederick, MD, USA). Labeled cDNAs were subjected to microarray gene expression profiling with The Cancer PathwayFinder GEArray (SuperArray Bioscience Corporation), a focused microarray with the sequences of 96 genes that control the cell cycle, apoptosis, growth factor signaling, angiogenesis, stem cells, cell proliferation, and tumor invasion and metastasis (27, 31). Array images were captured and digitized using the FluorChem 8800 Imaging System (Alpha Innotech, San Leandro, CA, USA), and each pair of the datasets were exported to GEArrayAnalyzer (SuperArray Bioscience Corporation). For each gene spotted on the array, gene expression was considered present if the integrated density value (IDV) was 1.5-fold or more the median IDV of the array. Otherwise, expression was considered to be absent. For comparison of the relative expression levels of a given gene between paired samples, the minimum IDV was used for background subtraction, and subsequently the median value was used for normalization. The expression level of a given gene was considered higher in one sample when this gene is “present” in the sample and is 1-2 fold greater than that of the normalized IDV and that of its counterpart on the paired array. After the analysis, the top up-regulated genes identified in cell clusters overlying focally-disrupted capsules were subjected to further assessment. The ratio of the total number of up-regulated genes to the total number of down-regulated genes within the gene set in all informative cases (defined as cases containing measurable signals in all or over 85% of the paired samples) was statistically calculated using the ratio >2 as the standard for cell clusters overlying focally disrupted capsules. The results of cDNA gene expression profiling in top up-regulated genes were verified using the real-time RT-PCR method. As these studies were not funded by NIH, the microarray results were not required to be submitted to the national microarray databases.
Results
All negative controls were consistently devoid of distinct immunoreactivity, and the results were highly consistent in duplicates or triplicates. The sub-cellular localization and reactive cell types for each of the antibodies were consistent with those of published data and specifications of manufacturers. The basal, myoepithelial, and luminal cell populations were morphologically and immunohistochemically distinguishable in all cases. The basal or myoepithelial cell population is characterized by its spindle or elongated shape with a densely-stained nucleus, and expresses CK 34βE12 or SMA in its cytoplasm and p63 in the nucleus. The luminal cell population is characterized by its round or oval shape with a lightly stained nucleus harboring distinct nucleoli, and is devoid of CK 34βE12, SMA and p63 expression. Although CD16, CD56, FOX3P, mast cells and neutrophils are often seen within the tumor tissues, the vast majority (over 95%) of infiltrated immune cells physically associated with focally-disrupted capsules, basal, myoepithelial, and luminal cells belong to a population of lymphocytes. Thus, for the convenience of description, tumor-infiltrating immune cells have been collectively referred to as infiltrating lymphocytes and both prostate basal and breast myoepithelial cells are referred to as basal cells below.
Our study has identified several unique and clinically-relevant features that have not been previously reported or sufficiently addressed.
Presence of lymphocyte aggregates in normal or hyperplastic breast and prostate tissues. Lymphocyte aggregates were seen in some normal-appearing, and in pre-invasive breast and prostate epithelial clusters, in which many infiltrating lymphocytes were physically attached to the basal cell layers that were often focally disrupted or substantially attenuated (Figure 1). In normal or pre-invasive epithelial structures with a largely continuous basal cell layer, the vast majority of the epithelial cells are physically segregated from, even they are partially or completely surrounded by, infiltrating lymphocytes (Figure 1E-H). In sharp contrast, significantly fewer infiltrating lymphocytes were seen within the invasive tissue component, and over 98% of the invasive cancer cells had no direct physical contact with infiltrating lymphocytes. Out of a total of 154 lymphocyte aggregates detected near epithelial structures, 147 (95.5%) were associated with normal or pre-invasive epithelial structures with focally-disrupted capsules, and only seven (4.5%) were located at the invasive cancer component (p<0.01).
Lymphocyte infiltration at the site of capsule disruptions of normal or hyperplastic tissues. Under high magnification, the vast majority of infiltrating lymphocytes within the normal or hyperplastic lesions were located at the site of focally-disrupted capsules (Figure 2). Infiltrating lymphocytes were often lined at the basal cell layers, mimicking the basal cells. Nearly all basal cells physically associated with infiltrating lymphocytes appeared to have been either largely dissolved by the lymphocytes or displayed a wide variety of degeneration-related changes, including loss of expression of phenotypic markers, vacuolation, fragmentation, swelling, nuclear membrane breakage, chromatin condensation, atrophy, and necrosis (Figure 2, thin arrows). In sharp contrast, the vast majority of the luminal cells physically associated with infiltrating lymphocytes at the site of focally disrupted basal cell layers displayed no distinct signs of degeneration (Figure 2).
Lymphocyte infiltration at the site of focal capsule disruptions of in situ or PIN tissues. Similarly to those seen in normal-appearing and hyperplastic tissues, the vast majority of infiltrating lymphocytes seen in a subset of in situ breast cancer and high-grade prostatic intraepithelial neoplasia were also located at or near focally-disrupted basal cell layers. In some cases, infiltrating lymphocytes were arranged as a capsule-like structure, partially surrounding the epithelium (Figure 3E-H). Again, nearly all basal cells physically associated with infiltrating lymphocytes exhibited distinct signs of degeneration-related changes, while the vast majority of luminal cells showed no distinct signs of degeneration (Figure 3E-H). Out of 876 basal cells physically associated with infiltrating lymphocytes, 792 (90%) showed cytological signs of degeneration. In contrast, of 474 luminal cells physically associated with infiltrating lymphocytes, only 59 (12%) had cytological signs of degeneration (p<0.01).
Elevated apoptosis in focally-disrupted basal cell layers. In both breast and prostate, over 85% of apoptotic cells were seen in the basal cell population. As shown in Figure 4, these apoptotic cells were exclusively located at or near the site of focally-disrupted basal cell layers. The vast majority of these apoptotic cells were apparently basal cells, judging from their location, shape, and the presence of apoptotic signals (thick arrows Figure 4) in the nucleus and CK 34βE12 signals (thin arrows Figure 4) in the cytoplasm of the same cell.
Significant reduction of tumor suppressor p63 expression in focally-disrupted basal cell layers. In sections double-immunostained for CK 34βE12 and p63, the vast majority of the basal cells in non-disrupted layers expressed both CK 34βE12 (red color in the cytoplasm) and p63 (black color in the nucleus). In contract, most basal cells in focally disrupted layers only expressed CK 34βE12, with the loss of morphological detail, swelling, and necrosis (Figure 5).
The absence of tumor suppressor p63 in degenerated basal cells is more easily appreciable in fluorescent double-immunostained sections. As shown in Figure 6 (which shows a set of three consecutive sections immunostained for CK 34βE12 and p63), a normal-appearing duct harbors a small focal disruption in its basal cell layer (circle), in which two morphologically-distinct basal cells near the disruption appear only to express CK 34βE12 (arrows), while most basal cells distant from the disruption express both CK 34βE12 and p63 (squares).
Elevated proliferation in some epithelial cell clusters surrounded by lymphocytes. In contrary to the belief that the physical contact between tumor cells and lymphocytes may lead to destruction of the associated tumor cells, some isolated epithelial cell clusters completely surrounded by infiltrating lymphocytes had a substantial higher proliferative index than their morphologically similar counterparts without surrounding lymphocytes. Figure 7 (a set of two consecutive sections double-immunostained for different markers) shows all cells of an isolated epithelial cell cluster surrounded by lymphocytes to express high levels of Ki-67, a cell proliferation-specific marker.
Elevated proliferation in cells overlying focally disrupted basal cell layers. In contrary to the degenerative changes seen in basal cells near focally-disrupted capsules, epithelial cells overlying focally-disrupted capsules often had a substantially higher proliferative index than their morphologically similar counterpart distant from the disruption. Figure 8 shows clusters of multiple Ki-67-positive cells exclusively seen at the site of focally-disrupted basal cell layers.
Cell budding from focally-disrupted tumor capsules. Focal tumor capsule disruptions coupled with lymphocyte infiltration appeared to facilitate proliferation of the overlying luminal cells. Of a majority of tumor nests with a large focal disruption (greater than 25% of the entire tumor capsule), the overlying luminal cells were often arranged as finger- or tongue-like projections invading into the stroma. In sections double-immunostained for CK AE1/AE3 and LCA, the tips of these projections were completely or partially surrounded by lymphocytes. In the immediately adjacent sections double-immunostained for CK AE1/AE3 and D2-40 (a specific marker for lymphatic ducts), a significant number of luminal cells were found to be dissociated from the tumor core, and some of them were at a distance from the tumor (Figure 9).
Elevated expression of proliferation- and stem cell-related genes. Gene expression profiling revealed that microdissected luminal cells overlying focal capsule disruptions had over 47-fold higher expression of LIF, a growth factor (Figure 10), compared to their adjacent counterparts distant from the site of the disruption. Microdissected luminal cells overlying focal capsule disruptions also had a more than 5-fold higher expression of two stem cell lineage markers, KIT and NCOR2 (Figure 10), compared to their adjacent counterparts distant from the site of the disruption. In addition, the expression levels of two endothelial cell markers (ENG and CAM2, two cytokines (CX3CL2 and CXCR4, and a TNF receptor (TNFRSF10D) were 6.38-, 12.12-, 6.14-, 12.81-, and 8.20-fold higher, resectively, than their counterparts at a distance from the site of focal capsule disruptions (Figure 10). It is interesting to note that the expression level of MMP-26, a major member of the proteolytic enzyme family, was 6.94-fold lower in luminal cells overlying focal capsule disruptions, compared to their counterparts distant from the site of the disruptions (Figure 10).
Lymphocyte aggregates in normal or pre-invasive breast and prostate tissues. Paraffin-embedded human breast (A-D) and prostate (E-H) tissue sections were double-immunostained for smooth muscle actin (SMA, a myoepithelial marker; red) plus leukocyte common antigen (LCA; brown), or CK 34βE12 (a basal cell marker; red) plus LCA. Circles identify lymphocyte aggregates surrounding epithelial structures with focally disrupted capsules. Asterisks identify the invasive tissue component. Thick and thin black arrows identify the myoepithelial or basal cell layer and associated immune cells, respectively. Thick and thin green arrows identify invasive cancer cells and associated immune cells, respectively. Note that lymphocyte aggregates are exclusively associated with normal or hyperplastic epithelial structures and the invasive component has significantly fewer infiltrating immune cells. Note that the vast majority of luminal cells within normal or pre-invasive epithelial structures are also shielded from immune cells by the basal cell layer. A, C, E, and G: ×100; B, D, F, and H: higher magnification (×300) of A, C, E, and G, respectively.
Lymphocyte infiltration at sites of focal capsule disruptions of normal-appearing tissues. Paraffin-embedded human breast (A-D) and prostate (E-H) tissue sections were double-immunostained for smooth muscle actin (SMA, a myoepithelial marker; red) plus leukocyte common antigen (LCA; brown) or CK 34βE12 (a basal cell marker; red) plus LCA. Circles in A, C, E and G identify structures shown at higher magnification in B, D, F, and H, respectively. Thick and thin black arrows identify the myoepithelial or basal cell layer and associated immune cells, respectively. Thick and thin green arrows identify luminal cells and associated immune cells, respectively. Note that the vast majority of the basal cells physically associated with immune cells show cytological signs of degeneration, whereas the vast majority of the luminal cells immediately adjacent to or physically associated with immune cells are cytologically-similar to their counterparts distant from immune cells. A, C, E, and G: ×200. B, D, F, and H: ×600.
Lymphocyte infiltration at the site of focal capsule disruptions of pre-invasive tissues. Paraffin-embedded human breast (A-D) and prostate (E-H) tissue sections were double-immunostained for smooth muscle actin (SMA, a myoepithelial marker; red) plus leukocyte common antigen (LCA; brown) or CK 34βE12 (a basal cell marker; red) plus LCA. Circles in A, C, E and G identify structures shown at higher magnification in B, D, F, and H, respectively. Thick and thin black arrows identify the myoepithelial or basal cell layer and associated immune cells, respectively. Thick and thin green arrows identify luminal cells and associated immune cells, respectively. Note that the vast majority of the basal cells physically associated with immune cells show cytological signs of degeneration, whereas the vast majority of the luminal cells immediately adjacent to or physically associated with immune cells are cytologically similar to their counterparts distant from immune cells. A, C, E, and G: ×200; B, D, F, and H: ×600.
Elevated apoptosis in focally-disrupted basal cell layers. Human breast (A-B) and prostate tissue sections (C-D) were subjected to apoptosis detection with an ApopTag® Plus Peroxidase In Situ Apoptosis Detection Kit. After the apoptosis assay, the sections were subjected to immunohistochemical staining for SMA or CK 34βE12 to determine the histological origin of the apoptotic cells. Circles in A and C identify structures shown at higher magnification in B and D. Arrows identify apoptotic cells. Note that all these apoptotic cells appear to be basal cells based on their location and expression of the myoepithelial or basal cell phenotypic markers. A and C: ×200; B and D: ×500.
Loss of p63 expression in basal cells detected by chromogranin-based immunostaining. Human breast (A-B) and prostate (C-D) tissue sections were double immunostained for SMA plus p63 or CK 34βE12 plus p63. Circles in A and C identify structures shown at higher magnification in B and D, respectively. Thick arrows identify cells with both SMA or 34βE12 and p63 expression. Thin arrows identify cells lacking p63 expression. Note that the vast majority of basal cells in largely continuous basal cell layers express both markers, while the majority of cells in focally-disrupted layers lack p63 expression. A and C: ×200; B and D: ×500.
Loss of p63 expression in basal cells detected by fluorescence-based immunostaining. A set of three adjacent human prostate tissue sections were double-immunostained for CK 34βE12 and p63, and the antigen–antibody complexes were distinguished with different chromogens (A, B) or using secondary antibodies labeled with different fluorophores (C-H). Circles identify focal basal cell layer disruptions. Squares identify basal cells distant from focal capsule disruption. Arrows identify two basal cells that express CK 34βE12 but lack p63 expression. Note that most basal cells distant from the site of focal disruption express both p63 and CK 34βE12. Original magnification, ×300.
Elevated proliferation in epithelial cell clusters surrounded by lymphocytes. A set of two consecutive human breast tissue sections were double-immunostained for SMA plus LCA or SMA plus Ki-67, respectively. Circles in A and C identify structures shown at higher magnification in B and D. Arrows identify Ki-67-positive cells surrounded by infiltrating immune cells. Note that these epithelial cells surrounded by infiltrating immune cells not only lack signs of degeneration, but also express high levels of the cell proliferation marker Ki-67. A and C: × 200; B and D: ×500.
Elevated proliferation in cells overlying focally-disrupted basal cell layers. Human prostate tissue sections were double-immunostained for CK 34βE12 and Ki-67. Black circles in A and C identify structures shown at higher magnification in B and D, respectively. Green circles identify Ki-67-positive cells overlying focal disruptions of the basal cell layers. A and C: ×200; B and D: ×500.
Luminal cell budding from focally-disrupted tumor capsules. Two sets of two immediate adjacent tissue sections of human breast (A-D) and prostate (E-H) tumors were double-immunostained for different markers. Circles in A, C, E and G identify structures shown at higher magnification in B, D, F, and H. Thick arrows identify ‘budding’ or dissociated tumor cells within lymphocyte aggregates. Thin arrows identify infiltrating immune cells in B and F, and lymphatic ducts in D and H. Note that tumor cell budding and dissociation are exclusively seen at the site with intensive immune cell infiltration and most budding or dissociated cells are physically associated with infiltrating lymphocytes. Note that the density of lymphatic ducts is also substantially higher within lymphocyte aggregates than in other sites. A, C, E, and G: × 150; B, D, F, and H: ×600.
Elevated expression of stem cell-associated genes in cells overlying focal basal cell layer disruptions. Circles identify cells microdissected for gene expression profiling and differentially expressed genes between cells overlying focal capsule disruptions and adjacent cells within the same epithelial structure but distant from the site of disruption. The table lists the names of differentially expressed genes and their relative levels.
Discussion
Our study of breast and prostate tumors revealed a number of unique features that have not been previously reported nor sufficiently analyzed. These include: i) all 154 lymphocyte aggregates detected were preferentially distributed within pre-invasive structures with focally-disrupted capsules; ii) infiltrating lymphocytes were predominantly seen at the site of focally-disrupted tumor capsules; iii) over 90% of the basal cells immediately abjacent to or physically associated with lymphocytes displayed a wide variety of cytological signs of degeneration, while only 12% of the luminal cells exhibited similar signs; iv) focal epithelial capsule disruptions coupled with lymphocyte infiltration appear to facilitate proliferation and dissociation of the overlying luminal cells. These findings are consistent with our hypothesis tha tumor-infiltrating lymphocytes are preferentially associated with degenerated basal cells but are rarely associated with the luminal cells at the initial or early stage of tumor invasion. To the best of our knowledge, this is the first report that demonstrates differential impact of infiltrating immune cells on the basal and luminal cells of human breast and prostate tumors.
The primary reason for the preferential association of tumor-infiltrating lymphocytes with the basal cell population is unknown, but appears to result from focal degeneration and disruption of the basal cell layers. Our previous studies have consistently shown that compared to its morphologically clear-cut and non-disrupted counterpart, a focally disrupted basal cell layer has a significantly lower expression of tumor suppressor p63 and proliferating cell nuclear antigen, but a significantly higher frequency of apoptosis and degeneration (26-37). As shown in Figure 5, the vast majority of basal cells within non-disrupted layers expressed both CK 34βE12 and p63, but most basal cells within disrupted layers lacked p63 expression and had distinct signs of degeneration, including the loss of morphological detail, fragmentation, swelling, nuclear membrane breakage, and necrosis. Thus, it is very likely that the basal cells in these patients may belong to an ‘aged’ or degenerated population. The degradation products of these basal cells may function as self-epitopes to attract immune cells or to stimulate the production of auto-antibodies. Consistent with our speculation is the fact that our recent study revealed that protease-degraded collagen I fragments of the breast tumor capsule function as a specific mediators attracting macrophage and other immune cell infiltration (46).
The breast and prostate epithelium is normally devoid of both blood vessels and lymphatic ducts, and the myoepithelial or basal cell layer is the sole source of several tumor suppressors (47-50). Thus, a focal degeneration or disruption in a given basal cell layer could lead to several focal alterations with significant consequences, including: i) localized loss of tumor suppressors and paracrine inhibitory functions, which confer epithelial cell growth advantages allowing these cells to escape from programmed cell death (51-55); ii) localized increase of permeability for oxygen, nutrients, and growth factors, which selectively favors the proliferation of progenitor or stem cells (56-58); iii) localized increase of infiltration of lymphocytes, which directly export growth factors to the epithelial cells through direct physical contact (59-64); iv) direct epithelial–stromal cell contact, which augments the expression of stromal matrix metalloproteinase or represses the normal production and distribution of E-cadherin and other cell adhesion molecules, facilitating epithelial to mesenchymal transition and cell motility (65-67); vi) direct exposure of the epithelial cells to different cytokines, which facilitates vasculogenic mimicry and tumor angiogenesis (68, 69); and vii) direct physical contact between newly formed cell clusters and stromal cells, which stimulates the production of tenascin and other invasion-associated molecules, facilitating stromal tissue remodeling and angiogenesis, providing a favorable microenvironment for epithelial cell migration and proliferation (70).
More importantly, as the stem cell population is rested at the basal cell layer, these alterations selectively favor the proliferation of progenitor or stem cells overlying focally-disrupted basal cell layers, or activate mitogen-protein kinases and protein kinase C that trigger the exit of the residual primitive stem cells from quiescence (54, 55). Thus, tumor cell clusters overlying focally disrupted capsules are very likely to represent a population of tumor progenitors with a greater propensity to progress to invasive or metastatic cancer. Consistent with our speculation is the fact that the expression levels of two stem cell lineage markers, KIT and NCOR2, were 5-fold higher in epithelial cells overlying focally disrupted capsules compared to cells within the same structure but distant from the region of focal disruptions (30, 31).
Since the disruption of the tumor capsule is an absolute prerequisite for tumor invasion and metastasis, our hypothesis, if confirmed, is likely to have significant scientific and clinical implications and applications. Scientifically, it may open a new paradigm for revealing the intrinsic mechanism of tumor invasion and metastasis. Clinically, as the molecular profile not only defines the scope and extent of risk of aggressive tumor biology, but also precedes morphological and biochemical changes, the application of our novel concept and technical approaches may lead to the identification of specific molecules attracting lymphocyte infiltration, which could potentially lead to the development of therapeutic agents to fortify the tumor capsule or to prevent its disruptions. Furthermore, the elucidation of the molecular profiles of the cell population overlying focal capsule disruptions may lead to the identification of clinically-relevant surrogate biomarkers that can be used to distinguish clinically aggressive and indolent pre-invasive tumors. In addition, our findings could offer a reasonable explanation for the contradictory reports and statements regarding the impact and clinical significance of immune cell infiltration into tumor tissues. Unfortunately, we do not have clinical follow-up data of the patients, as our cases are from different states of the USA and around the world, which makes clinical follow-up difficult, if not impossible. Therefore, our study cannot directly link our findings to metastasis-free survival or prognosis.
Acknowledgements
This study was supported, in part, by research grants DAMD17-01-1-0129, DAMD17-01-1-0130, PC051308 from Congressional Medical Research Programs; BCTR0706983 from the Susan G. Komen Breast Cancer Foundation; 05AA from AFIP/ARP joint research initiative projects; 2008-02 from US Military Cancer Institute and Henry M. Jackson Foundation to Dr. Yan-gao Man; and 2006CB910505 from the Ministry of Chinese Science and Technology Department to Drs. Xichen Zhang, Yan-gao Man, and Guiyuan Li.
Footnotes
-
Conflicts of Interest
The Authors declare that they have no conflicts of interest.
- Received June 24, 2014.
- Revision received July 29, 2014.
- Accepted August 4, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.