Mucositis A Case of Allostatic Load 231 Oral Mucositis

An important example of Type 2 allostasis in oral biology and medicine, which could benefit from fractal analysis, is the clinical condition of oropharyngeal mucositis (OM, Fig. 2.2). The term "mucositis" and "stomatitis" are often used interchangeably, but OM specifically pertains to pharyngeal-esophago-gastro-intestinal inflammation, that manifests as red, burn-like sore or ulcerations throughout the mouth.

Stomatitis is an inflammation of the oral tissues proper, which can present with or without sores, and is made worse by poor dental hygiene. Typically, stomatitis manifests 1 to 10 mm peripheral to a pathological lesion of the oral mucosa (Chiappelli 2005; Duncan and Grant 2003; Weissman, Janjan, and Byhardt 1989; Janjan, Weissman, and Pahule 1992).

Figure 2.2. Oral Mucositis—Pathological manifestation (intraoral photograph)—soft palatal and lingual lesions.

OM is a common and treatment-limiting side effect of cancer therapy, which affect up to 100% of patients undergoing high-dose chemotherapy and hematopoietic stem cell transplantation, 80% of patients with malignancies of the head and neck receiving radiotherapy, and over 50% of patients receiving systemic chemotherapy. It is a critical condition since in 2004 alone, close to 1.5 million Americans were diagnosed with cancer and received some combination therapies for this disease, including surgery, radiation, chemotherapy, immunotherapy and/or cell transplantation. OM is the principal oral complication from cancer therapies. In most cancer patients, OM manifests as clinically observed atrophy (tissue damage) and telangiectasis (blood vessel, spider-like red spots) of the oral mucosa, with significant increase the risk for pain and necrosis (Chiappelli 2005; Duncan and Grant 2003).

OM is a limiting factor in the effectiveness of cancer therapy because it leads to significant impairment in quality of life (Weissman et al. 1989; Janjan et al. 1992). Patients with OM often report impaired eating, drinking, swallowing, and talking, as well as disturbed sleep. Increased predisposition to local and systemic infections is also serious sequelae commonly associated with OM (Cengiz, Ozyar, Ozturk, Akyol, Atahan, and Hayran 1999). Dry mouth, mouth sores and pain, taste changes, and sore throat are among the most frequently reported troublesome or debilitating side effects of OM. Together, they lead to significant weight loss, overall drop in energy level, fatigue, lethargy and muscle weakness (Rose-Ped, Bellm, Epstein, Trotti, Gwede, and Fuchs 2002). All in all, OM can be so severe as to lead to the need to interrupt or discontinue the cancer therapy protocol, and to hinder cure of the primary disease (Chiappelli 2005).

An important caveat to these reports, however, must be noted: many of the signs and symptoms listed above are clinically observed as well in many cancer patients who undergo radiation treatment alone or in combination with chemotherapy, and who do not manifest overt signs of OM. Therefore, it is often held that OM is one facet of the complex spectrum of side effects of cancer intervention, rather than the cause per se of the symptomatology described above.

What is certain is that OM is an acute oral mucosal inflammatory reaction secondary to cell death of the basal epithelial cell lining. The oral mucosa is a rapidly replacing body tissue: one that has received little attention in terms of defining its cell kinetics and cellular organization, relative to other domains of cell biology. The tissue is sensitive to the effects of cytotoxic agents, the consequence of which can be progenitor cell death, with the subsequent development of the symptoms of cell death and, consequentially, of OM. Radiation and chemotherapeutic agents act by blocking mitosis of rapidly dividing cell, including cellular toxicity and apoptotic cell death, and by decreasing their regeneration. These effectors kill cancer cells and fast-proliferating normal cells, such as oro-gastro-intestinal epithelial cells, thus generating lesions. Immune cells are henceforth actively recruited to migrate at the site of these emerging lesions, and an inflammatory state and alteration of oral flora ensue. An inflammatory lesion is what is commonly recognized as OM. The involvement of immunity in OM can involve systemic immune suppression, significantly increased risk of local and systemic infection, opportunistic infections and mortality due to sepsis. In brief, the loss of rapidly dividing epithelial progenitor cells triggers the onset of OM, but its severity and duration are determined by changes in local oral immunity (Chiappelli 2005; Duncan and Grant 2003; Potten et al. 2002; Alvarado, Bellm, and Giles 2002; Blijlevens, Donnelly, and DePauw 2005).

Oral microorganisms play an important role in aggravating the pathology of the impaired epithelium, and individual oral hygiene has considerable influence on symptom presence and severity. Common dental care products that offer several different active ingredients to reduce bacterial counts and plaque, such as fluorides, have antibacterial (sodium fluoride, amine fluoride, tin fluoride) and plaque decreasing effects (amine fluoride and tin fluoride), and substantially improve the prognosis of OM (Chiappelli 2005). Chlorhexidine is antimicrobial and very effective in inhibiting plaque growth, but, because of side effects, which include mucosal irritation, is not recommended as a viable for long-term use for OM. Smoking, and alcohol use and abuse, and psychoemotional stress are reported to contribute to exacerbate OM (Chiappelli 2005; Janjan et al. 1992; Cengiz et al. 1999; Alvarado et al. 2002; Kwong 2004).

Certain cytokines play an important role in OM. Evidence-based research has established that applications of granulocyte-macrophage colony-stimulating factor (GM-CSF) plays a critical role in preventing the onset or the exacerbation of oropharyngeal mucositis (risk ratio, RR = 0.51, CI95: 0.29-0.91; NNT = 3, CI95: 2-20) (Clarkson, Worthington, and Eden 2003). Clinical trials suggest that GM-CSF has clinical benefits beyond enhancing neutrophil recovery, including shortening the duration of mucositis and diarrhea (Nemunaitis et al. 1995), stimulating dendritic cells, preventing infection, acting as an adjuvant vaccine agent, and facilitating antitumor activity (Buchsel, Forgey, Grape, and Hamann 2002). GM-CSF belongs to the family of colony-stimulating factors (CSFs), which are central to the development and maturation of cells of the immune system, the modulation of their functional responses, as well as the maintenance of immune homeostasis and overall immunity. This group of glycoproteins consists of the macrophage-CSF (M-CSF), granulocyte-CSF (G-CSF), granulocyte/macrophage-CSF (GM-CSF), and multi-CSF (interleukin[IL]-3). GM-CSF functions at early stages of lineage commitment regulating the expansion and maturation of hematopoietic progenitor cells (Barreda, Hanington, and Belosevic 2004; Chiappelli in press).

GM-CSF is a proinflammatory cytokine, which stimulates proliferation and differentiation of neutrophilic, eosinophilic and monocytic lineages of cellular immunity. In immunosuppressed patients, and in murine models of therapeutic immune suppression, GM-CSF administration is effective in boosting the innate immune response, while continuing to suppress the adaptive immune response to prevent graft rejection (Chiappelli 2005; Xu, Lucas, and Wendel 2004). GM-CSF belongs to a pattern of cytokines distinct to the traditionally recognized TH1 and TH2 patterns. By its action on dendritic cells, GM-CSF modulates the production of IL-23 by these cells. IL-23 belongs to the IL-12 cytokine family, which shares p40, and binds to the IL12P1/IL-23R receptor complex to play a critical role in end-stage inflammation via signaling of the JAK/STAT pathway. Administration of IL-11 down-regulates IL-12. IL-23 augments production of IL-17 by T cells, which becomes the significant cytokine in the involvement of T cells in inflammation. IL-17 leads to decreased neutrophilia in animal models. In vitro administration of IL-17 synergizes the effects of TNF-a for GM-CSF induction, for increased intercellular adhesion molecule (ICAM)-1 (CD54) expression by CD34+ progenitors, for increased neutrophil maturation, and for increased proinflammatory IL-6 and prostaglandin(PG)E2 and G-CSF growth factor production by keratinocytes and endothelial cells. IL-17 also favors increased secretion of the migration-inducing cytokine, IL-8. In human U937 cells, and probably in all lymphoid and myeloid cells, the JAK/STAT signaling pathway is responsible for transducing signals from the IL-17 receptor. Human vascular embryonic cells (HUVEC) express a homologue of IL-17R, as do ductal epithelial cells of human salivary glands, suggesting a significant role for IL-17 in immunity in the oral mucosa. In fact, in inflamed gingival tissue, IL-17 is the predominant cytokine in tissue proximal to 4 to 5 mm diseased periodontal pockets. IL-17 plays a critical role in immunity and mucosal inflammation, by modulating the effects of GM-CSF and other growth factors (Watford, Hissong, Bream, Kann, Muul, and O'Shea 2004; Chiappelli in press).

In an evidence-based research study on the comparison of utilization of complementary and alternative (CAM) treatment for OM to traditional pharmacological intervention (e.g., GM-CSF), we have examined 424 reports. Critical analysis suggests that traditional medical intervention is not inferior to CAM in the treatment of mucositis. Issues in the quality of the research methodology were identified that hamper the overall quality of the data, including the inadequate use of double blinding, randomization, and placebo-controlled, and the lack of satisfactory statistical analysis of the data. Research protocols in this area often suffer from significant weaknesses, which hamper their efficacy. No meta-analysis was generated because of the nonhomogeneity of measures, and the lack of statistical stringency.

2.3.2 Neuroendocrine-Immune Orderly Chaos in OM

The immune system is a sensory organ for external stimuli, such as local cellular trauma caused by histo- and cyto-pathology and infectious agents that cannot be detected directly by the nervous system (vide supra). The bidirectional nature of the communication between the immune and nervous systems is responsible for directing the organism's response to infection and to inflammation (Solomon 1987; Chrousos and Gold 1992; Kiecolt-Glaser et al. 2002; Irwin 2002; Chiappelli et al. in press-c; Engle 1960; Prolo and Chiappelli in press; Chiappelli et al. 2004). The physiological challenge that ensues following sensory perceptions, including sympathetic, parasympathetic and hormonal responses, plays a crucial role in regulating neuroendocrine-immunity, thus ensuring the functionality and the relevance of the psychoneuroimmune loop. Because both the nervous and the immune systems have the ability to learn, and the capacity for memory, and because cells of both systems are endowed with receptors for neural and immune factors, the lymphocyte was termed the "mobile brain" (Pierpaoli and Maetroni 1988).

The intercellular communication between the nervous and immune systems via common receptors and signal molecules yields a conceptual framework for this cross talk. Soluble immune regulatory factors, including cytokines (e.g., proinflammatory, TH1, TH2), growth factors (e.g., GM-CSF) modulate neurobiological responses at rest and during the allostatic response. Neurotransmitters (e.g., epinephrine, acetylcholine), neuropeptides (e.g., endogenous opioids), and hormones (e.g., glucocorticoids) have specific receptors on and in cells of the immune system, that contribute to mediating immune events involved in inflammation, surveillance and healing (i.e., during allostatic load and overload). This cross talk has important implications in maintaining the health of tissues, including the oral mucosa (Chiappelli et al. in press-c; Chiappelli in press; Blalock 1994). "Hardware" connections exist between the nervous system and immune organs (Bulloch, Cullen, Schwartz, and Longo 1987; Felten et al. 1987) that are relevant to the physiological cross talk between the psychoneuroendocrine and the immune system (Sterling and Eyer 1988; McEwen and Wingfield 2003; Chiappelli and Cajulis 2004; Chiappelli et al. in press-c; Chiappelli et al. 2004; Besedovsky and del Rey 1996).

It is possible and even probable that the severity of tissue damage in OM, as well as the pattern and rate of recovery of the oral mucosa depend in large part on its neurobiological structure. The oral mucosa is richly innervated by sympathetic and parasympathetic fibers, which establish an intimate dialogue with resident and invading myeloid and lymphoid cells in the healthy tissue as well as in pathological oral lesions (e.g., OLP, 41). Our studies of the glossopharyngeal nerve demonstrated the implications of the psychoneuroendocrine-immune network within the oral mucosa (Romeo, Tio, Rahman, Chiappelli, and Taylor 2001).

As noted above, fractal analysis has emerged in the biological sciences in the last decade to provide a new powerful look into physiological processes that follow nonlinear dynamics, as well as a reliable approach to quantify the complexity of cells and cellular components and organelles. Since function is interdependent with structure, fractal analysis, by elucidating the fundamental morphological changes in immune cells at the cellular and the nuclear levels following psychoneuroendocrine modulation, will contribute to our mounting understanding and knowledge of immunophysiology. The fractal characterization of human T lymphocytes, with emphasis on the Tregs subpopulation will be critical in a variety of clinical situations, including oral pathology (Chiappelli et al. 2005), and OM in particular.

The analysis of fractal dimension yields a noninteger measure of how "complicated" a self-similar object is, and is strictly greater than the integer derived from a topological Euclidean dimension.3 Fractals have a noninteger dimension, which represents the measure of space-filling ability of the object under study. The fractal dimension allows the quantitative comparison and categorization of fractals. By contrast, the traditional Euclidean dimension is the number of coordinates required to specify the object. A straight line, for example, has a Euclidean dimension of 1, but a plane has an Euclidean dimension of 2. A jagged line takes up more space than a straight line, and has a fractal dimension between 1 and 2 (e.g., 1.56). Fractal dimensions are, therefore, usually, albeit not always, about 20% larger than Euclidean dimensions, and are all the more informative, in that they represent the complexity of the object under measurement (e.g., T lymphocytes invading the oral mucosa in OLP, Chiappelli et al. 2005).

In the traditional sense of the word, "dimension" refers to a measure of how an object fills space, a topological space. This measure, the Lebesgue (Henri Lebesgue, 1875-1941) covering dimension, is the topological dimension, and is rendered by the archetypal example of the Euclidean «-space with a topological dimension n. (Lebesgue H. 1904. Leçons sur l'Intégration et la Recherche des Fonctions Primitives, Professées au Collège de France. Paris, Gauthier-Villars).

Lacunarity analysis is another a multiscaled method of this approach to object analysis, typically a sliding box algorithm, for describing patterns of spatial dispersion, which is used with both binary and quantitative data in one, two, and three dimensions. It is a parsimonious analysis of the overall fraction of a map, or transect, covered by the attribute of interest, the presence of self-similarity, the presence and scale of randomness, and the existence of hierarchical structure. For self-similar objects such as cells, lacunarity analysis can be used to determine the fractal dimension by readily interpretable graphic results. Lacunarity, X, is a counterpart to fractal dimension in that it describes the texture of a fractal. In more practical terms, lacunarity is concerned with the size distribution of the holes that compose a fractal. High lacunarity pertains to a fractal that has large gaps or holes. A fractal that is translationally invariant has characteristically low lacunarity. Different objects could have the same fractal dimension, but appear largely different because they differ in lacunarity (Mandelbrot 1994).

The use of fractal dimension and lacunarity has increasingly permeated the biological sciences. Case in point, fractal dimension was shown to be a reliable measure of atypical nuclei in dysplastic lesions of the cervix uteri (Sedivy, Windischberger, Svozil, Moser, and Breitenecker 1999). In that study, fractal dimension was obtained by the box-counting method. This approach involves starting with an image of which the fractal dimension is sought. The image is translated into a binary version where the pixels above a given cut off brightness are set to one, and the others are set to zero. An image of contours around the bright sections is generated, and divided up into boxes of a given size. The count of the number of boxes required to contain the contour is obtained, and the plot of the log of the box size vs. the numbers obtained for that box size provides the best-fit slope estimate, which corresponds to the fractal dimension of the image. That study impressed the notion that whereas the same fractal dimension could characterize certain nuclei, they differed substantially in their porosity, their lacunarity (Sedivy et al. 1999).

Box-counting fractal analysis was used to compare the hepatocyte nuclei obtained from normal liver and from hepatocellular carcinoma. In that study, the number of grid squares (boxes, tiles) that the cell border-image contacts, regardless of the number of pixels contained within the border, was counted. The log number of squares was plotted against the log of the box edge length, rendered as

where DB represented the box-counting fractal dimension of the nuclear image in binary form, and N(e) the smallest number of boxes with side length e that completely covered the outline of the object under study. The log-log plot of the (1/£*) (X-axis) versus N(e) (Y-axis) was derived to render the linear approximation of the first-degree polynomial equation (Y = a + mX). The slope of the least square fitted line, m, thus obtained is taken to represent DB, the box-counting fractal dimension. Self-similarity can be established by means of the evaluation of the fit of regression lines. Higher values for the coefficient of determination are representative of better fit. The closer the coefficient of determination is to 1.0, the larger the proportion of random error accounted for, and the greater the extent of self-similarity of the objects under study (Kerenji, Bozovic, Tasic, Budimlija, Klem, and Polozovic 2000).

Common approaches of fractal analysis of self-similarity include, in addition to the box analysis, the Hurst4 exponent. This value measures the smoothness of fractal time series and is based on the asymptotic behavior of the rescaled range of the process, the perimeter-area dimension analysis, the information dimension analysis, the mass dimension analysis, and the ruler dimension analysis. Common fractal dimension calculation methods for self-affine traces can also include rescaled range analysis, power-spectral analysis, roughness length analysis, variogram, and wavelets. Fractal dimension of size-frequency data is most commonly estimated by the method of fragmentation dimension. It is unquestionable that the assessment of the fractal dimension has become a useful tool in quantitative histology and cytology over the past decade. This measurement is reliably obtained from computerized image analysis systems. It is generally accurate, with errors of < 1.5%, and reproducible (reliability coefficient, r = 0.972, Cf5: 0.868-0.987) (Cross, Cotton, and Underwood 1994).

Fractal dimension analysis adequately distinguishes cell populations with complex plasma membranes at diverse stages of maturation. Oligodendrocytes in culture, for example, were analyzed by the box-counting method. In order to generate estimates on the ascending part of the curve, which encompasses the range in which cells express the property of self-similarity, box size was restricted between 1 and 20 ^M. Larger boxes provide no additional information on cell morphology, and smaller boxes are well within the lower-end sensitivity range (0.1 ^M) of the typical analysis software,5 but provide no further information about the cells filamentous projections, typically in the 1 ^M range. Below this range, cells typically loose the property of self-similarity6 (Takeda, Ishikawa, Ohtomo, Kobayashi, and Matsuoka 1992; Bernard, Bossu, and Gaillard 2001). In brief, future studies aimed at providing lacunarity and box-counting fractal analysis assessments of invading lymphocyte subpopulations at different stages of OM pathology and treatment will be most revealing of OM immunopathology.

A principal caveat of this analytical protocol, however, lies in the fact that errors in fractal quantification arise because estimates are derived from pixel counts, a function of the number of pixels cut by the boundary of the object. Cellular contours with a fractal dimension greater than the planar Euclidean dimension of 1.0 will

British hydrologist, H.E. Hurst, who systematically characterized the hydrologic properties of the Nile (Hurst, H.E., Black, R.P. and Simaika, Y.M. 1965. Long-Term Storage: An Experimental Study. Constable, London).

e.g., Benoit, Trusoft

6 Prof. S. Gaillard, pers. commun.

appear more complex as the pixel size decreases and the resolution increases. By contrast, the more contorted boundaries will have higher fractal dimensions, with less error of estimate. Flawed interpretations of pixel-based quantification of fractal dimension also arise as a function of the distribution of the pixels: a compact distribution yields typically more reliable results and less error, compared to more widely dispersed, diffuse, or patchy pixel phenomenon. Typically, a scattered phenomenon yields the following relationship:

Scatteredness: (SD/dimension mean) x 100 = (dimension mean - 1.00)/2

Lymphocytes are typically smaller and less complex than oligodendrocytes. Nevertheless, the perimeters of the surface membranes of different lymphocyte populations can be digitized from electron microphotographs, and the data analyzed to derive the fractal perimeter dimension. Values range from 1.02 to 1.34, independently of the magnification, and are consistently obtained across the lymphocyte population of the same type (Chiappelli et al. 2005; Keough, Hyam, Pink, and Quinn 1991).

Figure 2.3. Fractal dimension in CD25+ and CD69+ human T lymphocytes (measured by box counting or by Euclidean distance map): oral lichen planus: fractal dimension in CD25+ and CD69+ human T lymphocytes measured by the box-counting method, compared to Euclidean distance dimension map. Biopsies from the oral mucosa from patients with oral lichen planus were obtained, fixed and process for standard immunohistochemistry. T cells at different stages of activation were identified with either monoclonal antihuman CD25 or antihuman CD69, and visualized by horseraddish peroxidase staining. Micrographs (60x + 10x eyepiece magnification), were obtained, and the stained cells isolated and adjusted the brightness by means of computer-aided software (Micrografx picture publisher). Euclidean estimates of size (i.e., relative estimates of cell radius) as well as fractal dimension were obtained by box-counting (i.e., relative complexity of the staining pattern on the cell membrane) with the FractalFox software (Qichang Li, Univ. Memphis).

Figure 2.3. Fractal dimension in CD25+ and CD69+ human T lymphocytes (measured by box counting or by Euclidean distance map): oral lichen planus: fractal dimension in CD25+ and CD69+ human T lymphocytes measured by the box-counting method, compared to Euclidean distance dimension map. Biopsies from the oral mucosa from patients with oral lichen planus were obtained, fixed and process for standard immunohistochemistry. T cells at different stages of activation were identified with either monoclonal antihuman CD25 or antihuman CD69, and visualized by horseraddish peroxidase staining. Micrographs (60x + 10x eyepiece magnification), were obtained, and the stained cells isolated and adjusted the brightness by means of computer-aided software (Micrografx picture publisher). Euclidean estimates of size (i.e., relative estimates of cell radius) as well as fractal dimension were obtained by box-counting (i.e., relative complexity of the staining pattern on the cell membrane) with the FractalFox software (Qichang Li, Univ. Memphis).

Immunological studies have confirmed that when blood cells are projected into an image plane their contours appear as borderlines of irregular shape with the property of the statistical self-similarity. The fractal perimeter dimension of normal peripheral blood mononuclear cells and mature T lymphocytes fall in the median range of 1.23 to 1.17, and the same dimension obtained from T cells of hairy-cell leukemia with highly convoluted morphology range between 1.32 and 1.36. By contrast, blast cells obtained of acute lymphoblastic leukemia and generated from activation of normal T cells in vitro appeared larger with a less complex membrane—stretched, as it were. They yielded lower fractal perimeter dimensions in the range of 1.11 to 1.16 (Losa, Baumann, and Nonnenmacher 1992). Human T lymphocytes from normal donors, and hairy leukemic cells exhibit distinct fractal dimension (mean ± SD: 1.15 ± 0.03 for normal T cells, compared to 1.34 ± 0.04 for hairy leukemic T cells, p < 0.05) (Nonnenmacher, Baumann, Barth, and Losa 1994). In our studies, active T lymphocytes that invade the oral mucosa in OLP have similar dimensions (Chiappelli et al. 2005), which suggests that immune cells that will be found in OM pathological tissue ought to be characterized by contours in that same range.

Specifically, our estimates of scatterdness, based on the relationship provided above, suggested an even distribution of the staining pattern around the circumference of the cell membrane, as opposed to compacted and clumped patterns of staining. Our data showed a reliability of measurement of 0.92 ± 0.11 (Fig. 2.3).

The data summarized in Fig. 2.3 show that the box-counting (fractal dimension) and Euclidean distance was not significantly different among CD69-stained cells, and was close to 1.0 (mean ± SD: 1.056 ± 0.029 and 1.029 ± 0.056, respectively, p > 0.05), suggesting that the configuration of these cells upon the slide was rather planar and linear. The fractal dimension of CD25-stained cells was significantly lower than their CD69 counterpart (0.934 ± 0.036 and 1.056 ± 0.29, respectively, p < 0.0001), suggesting that the membrane contours of CD25+ T cells was overall smoother than that of CD69+ T cells. By contrast, the Euclidean dimension of CD25-stained cells was significantly larger than that of CD69-stained cells (1.104 ± 0.043 and 1.029 ± 0.056, respectively, p < 0.028), indicating that CD25+ T cells were overall larger (larger relative estimate of cell radius), compared to CD69+ T cells.

One interpretation of these observations suggests the possibility that CD25+ T cells show signs of stretching of the plasma membrane as they enlarge toward the blast stage of cell activation. Euclidean dimension for both CD25+ and CD69+ T cells was freer of random error, with 98% of the variance accounted for. Box-counting assessments, by contrast, suffered from more random error across both cell populations with 75% of the variance accounted for. Such stretching is precisely what has been observed in cellular immune studies of T cell activation: early following antigenic trigger, T cells express CD69. Expression is rather transient, and leads to continued signaling pathways, which result in the expression of IL-2 and the a chain of its receptor, CD25 within a few hours. During this time, the cell advances through G1 and S of the cell cycle, and enlarges toward its blast state, which is fully attained several hours later (Chiappelli in press). In brief, fractal dimension analysis appears to be sensitive enough to distinguish CD69+ from CD25+ activated T cells.

It is possible and even probable, therefore, that this approach can be reliably utilized to identify subpopulations of Tregs.

Fractal dimension or lacunarity, as well as the fractal dimension of the epithelial-connective tissue, are associated with functional changes of transformed lymphocytic cells (Sedivy et al. 1999). The Zipf law7 can serve to characterize the T lymphocyte repertoire. In a study of murine CD8+ T cells, the fractal dimension derived from the Zipf plots correlated significantly (p > 0.05) with the nature of the repertoire (Burgos and Moreno-Tovar 1996). In a similar vein, distinct CD8+ cytotoxic T cell clonotypes were established on the basis of the unique DNA sequence of the third complementarity-determining region of the TcR P-chain. The frequency distribution of the clonotypes indicated a complex population pattern probably dependent upon mechanisms for maintaining a large number of antigen-specific clonotypes at a low frequency in the memory pool. The clonotypes were ranked in terms of population characteristics similar to power law distribution. When the repertoire was divided into specific subsets, based on their immunogenic properties, the resulting clonotype frequencies described a power law-like distribution, which indicates self-similarity of the repertoire, in which smaller pieces are slightly altered copies of the larger piece. The power law-like description was found to be stable in time, and was replicable across donors.

Self-similarity and power laws are associated with fractal systems, demonstrating the applicability of fractal analysis to the characterization of memory CD8 T cell repertoires (Naumov, Naumova, Hogan, Selin, and Gorski 2003), and could serve to characterize diverse repertoires of invading immune cells during the OM immunopathological process. This is particularly true and applicable to the Tregs subpopulation because Tregs easily migrate from the peripheral blood, to the lymphatic system, to the oro-pharyngo-gastro-intestinal mucosa (Chiappelli in press).

In brief, it is possible and even probable that fractal analysis will generate new insights about neuroendocrine-immune interaction in OM. Fractal analysis will also undoubtedly emerge as a novel and reliable experimental approach for the

Relations of the form fx) = k x are called power laws, but do not imply fractal structure. Power-laws imply that small occurrences are extremely common, whereas large instances are extremely rare, and are also referred to as Zipf or as Pareto relations. The Zipf law (George Kingsley Zipf, Harvard linguistics professor, 1920-1950) refers to the "size" y of an occurrence of an event relative to it's rank r, and seeks to determine the "size" of the 3rd or 8th or 100th most common occurrence. Size denotes the frequency of occurrence. Zipf's law states that the size of the rth largest occurrence of the event is inversely proportional to it's rank: y ~ rJ, with b close to unity. By contrast, Pareto's law (Vilfredo Pareto, Italian economist, 1948-1923) is given in terms of the cumulative distribution function, that is the number of events larger than x is an inverse power of x: P[X > x] ~ x~*. Pareto's distribution reports the number of events at any given x, and represents the probability distribution function derived from Pareto's law.

elucidation of the genomic, proteonomic and interactomic molecular cartography (vide infra) that defines and characterizes the phenotypic and the functional properties of Tregs and other immune cell subpopulations involved in OM and other stomatological pathologies.

2.3.3 Circadian Neurobiology and Neuroimmunopathology in OM

According to the guidelines for prophylaxis and therapy of mucositis ( [2005-05-04]), the risk to develop a grade 3 (severe) or grade 4 (life threatening) OM lies between 2 and 66%, depending on the type of chemotherapy (without additional radiotherapy), and between 1 and 53% depending on the type of tumor. In over 30% of the patients who develop a grade 3 or 4 OM, the start of the next chemotherapy cycle needs to be delayed, resulting in impaired cancer treatment (http://www3.interscience. [2005-05-04]).

Chemotherapeutic interventions that use 5-fluorouracil (FU) commonly lead to OM. The relationship between FU, one of the oldest and most widely used anticancer drugs, pharmacokinetics and patient response support a link between systemic drug exposure, physiology, and drug response, and consequentially survival. Hence, the maximal tolerated exposure is derived from pharmacokinetic follow-up and individual dose adjustment. Maximal tolerated exposure is dependent upon FU individual catabolic rate (Milano and Chamorey 2002).

The reported circadian pattern in epithelial cell response to insult following radiation or chemotherapy is likely consequential to neurobiological regulation, as well as biological circadian pattern inherent to the biochemical and molecular properties of certain populations of cells (e.g., expression of cell-autonomous Dec or Per genes). This hypothesis has barely been tested in isolated cell systems, and more in-depth and rigorously focused research is needed.

Data to date indicate that the activity of dihydropyrimidine dehydrogenase (DPD8), the rate-limiting enzyme in FU catabolism, varies according to the time of day in pattern reminiscent of the circadian variation of cortisol in normal subjects (Raida et al. 2002). This observation suggests that glucocorticoids and glucocorticoids-responsive elements in the nuclear compartment may mediate the regulation of the DPD gene. Whereas physiologic data from both human and animal investigations confirm circadian rhythmicity in DPD activity (Milano and Chamorey 2002), genomic and epigenetic (i.e., by noncoding DNA, vide infra) studies on the regulation of the DPD gene in the chromatin must now be designed.

These investigations will be particularly important because recent evidence indicates fine circadian regulation of DPD. For example, in an in vivo murine model of OM, the toxicity of FU on the oral mucosa was decreased by two to eightfold if

this drug was injected near the middle of the day (rest span) rather than in the middle of the night (activity span). Since the rhythm in FU toxicity appears to be linked to the sleep-wakefulness endogenous circadian cycle across species, the least toxic time in patients would correspond to 4:00 am (Levi et al. 1995). Furthermore, in an experimental model of neonatal male rats, DPD activity in liver increased 48 h following glucocorticoids injection (Fujimoto, Kikugawa, Kaneko, and Tamaki 1992).

Experiments in vitro have demonstrated that DPD activity is expressed in most tissues, including epithelial cells and leukocytes, and is highest in hepatocytes. Peripheral blood mononuclear cells (PBMC) are routinely used to monitor clinically DPD activity, and a significant, albeit weak, correlation between PBMC and liver DPD activity has been reported (Milano and Chamorey 2002). Among circulating leukocytes, myeloid derivatives (e.g., monocytes, macrophages) express greater activity than their lymphoid counterpart (i.e., T & B cells). A significant positive correlation was reported between PBMC-DPD activity and the differential percentage of monocytes/macrophages, which contributes to the observed inter- and intrapatient variability in the activity of DPD among the patients (van Kuilenburg, Meinsma, and van Gennip 2004).

Oral mucosa samples taken from healthy subjects show circadian variations in DPD activity. These findings were obtained by measuring DPD activity by HPLC. The data revealed a change in enzymatic activity from 0.004 to 0.13 nmol/min/mg protein at 10:00 h and from 0.07 to 0.16 nmol/min/mg protein at 24:00 h, with a 30% average relative increase from morning to midnight (range: -20% to +100%; p < 0.073). This pattern mirrors that of DPD activity observed in leukocytes. In brief, DPD rhythmicity in the oral mucosa cells peaks at 01:00 h, which is an important determinant of FU tolerability in the context of OM. These findings have considerable potential clinical relevance for the pattern of chronomodulated delivery of FU chemotherapy (Barrat, Renee, Mormont, Milano, and Levi 2003).

However, the pattern of circadian DPD activity levels can vary considerably, particularly among subjects. In some cancer patients, and for biological reasons that are not fully elucidated as of yet, this circadian pattern of DPD activity vanishes. The diurnal cortisol pattern showed the expected consistent circadian rhythm in the control subjects, and a blunted circadian cortisol pattern in a group of patients with advanced gastrointestinal carcinomas treated with FU. A trend towards a circadian rhythm of DPD mRNA expression in PBMC was also observed that corresponded roughly to the cortisol curve (i.e., peak of DPD mRNA expression at 05:00 h, trough at 14:00 h, p < 0.005 Mann-Whitney-Wilcoxon test). The DPD mRNA circadian pattern could be fitted to a cosine wave (p = 0.001, 0.014, Cosinor analysis) in 40% of the control subjects, but in not in the cancer patients. In brief, in these cancer patients, PBMC-DPD mRNA expression showed no trend toward consistent circadian rhythmicity, whereas circadian endogenous cortisol secretion pattern was maintained (Raida et al. 2002).

The principal caveats to the study of the circadian physiology of DPD in the context of OM lie in the fact that the relationship between PBMC-DPD activity and FU systemic clearance is weak (r = 0.10). Thus, simply determining PBMC-DPD is not sufficient to predict accurately FU clearance in general, and OM in particular. Evidently several as-of-yet undetermined intervening variables contribute to the maximal tolerated exposure. Case in point, population pharmacokinetic studies identify patient covariables that influence FU clearance, including increased age, high serum alkaline phosphatase, duration of drug infusion, in addition to low PBMC-DPD (Milano and Chamorey 2002).

The observations about the circadian pattern of DPD have brought forth the realization that optimal maximal tolerated exposure to FU could be obtained by carefully controlling the timing of FU administration in a chronomodulated mode. Research data overall support this hypothesis. Plasma pharmacokinetics of FU (600 mg/m2/day) differs among patients with advanced colorectal cancer, when the drug is administered with a programmable-in-time pump by continuous infusion for 5 days, either at a constant rate of delivery or with a chronomodulated rate. The peak of FU appears at 04:00 h in the chronomodulated schedule of administration. When FU is administered at a constant rate, mean plasma concentration varies in a circadian manner each treatment day, with a peak at 04:00 h (approximately 800 ng/ml) and a trough at 13:00 h (approximately 100 ng/ml). Severe OM was invariably observed by all patients on a flat schedule of FU chemotherapy, but only in a small proportion of patients on the chronomodulated schedule (p < 0.008). OM was generally less pronounced among patients with circadian rhythms in FU, which suggests that one mode of control of OM may be, in certain groups of cancer patients at least, via amplification or induction of FU circadian rhythms (Metzger et al. 1994).

In a related study, in patients receiving protracted low dose FU infusion, the circadian rhythm in FU plasma concentration was observed to peak at 11:00 h and to be lowest at 23:00 h, on average. The inverse relationship observed between the circadian profile of FU plasma concentration and PBMC-DP activity in these same patients indicated a link between DPD activity and FU pharmacokinetics. When the impact of the biological time of drug administration was also studied with short venous infusions; clearance was 70% greater at 13:00 h than at 01:00h. Similarly, peak drug concentration occurred in the first half of the night in patients receiving constant rate FU infusion for 2 to 5 days. Wide interindividual variations in the timing of the peak and trough of the 24 h rhythm in DPD activity were observed (Milano and Chamorey 2002).

Consequently, a rationale for FU chronomodulated therapy emerged based on the apparent circadian rhythm in host drug tolerance, which is greatest during the nighttime when the proliferation of normal target tissue is least. A randomized study of chronomodulated FU therapy with maximal delivery rate at 04:00 h further showed to be significantly more effective and less toxic than control flat FU therapy (Milano and Chamorey 2002). Other studies have shown mixed results of chronomodulated chemotherapy regimens with variable-rate infusions of FU have shown mixed results (Raida et al. 2002).

The novel tools of evidence-based research in medicine and of systematic review were employed to critically review the pertinent available research. We evaluated the evidence in support of the hypothesis that chronomodulated delivery of FU could be recommended, compared to control flat FU administration for therapeutic cancer intervention, in terms of avoiding the side effects of OM. The process for the search of the literature, critical evaluation and analysis of the evaluations were as described elsewhere (Chiappelli, Navarro, Moradi, Manfrini, and Prolo 2006a). The number needed top treat could not be computed because the design of the studies failed to include appropriate control groups, when considering the efficacy of a particular chemotherapy regimen. A X analysis was used to test if the difference between the percentage of patients who developed mucositis was statistically significant between the groups. From a total of 291 papers, 63 were included in this study, of which 33 implemented a continuous treatment of FU, and 30 studies reported chrono-modulated administration of FU. Of a total of 1395 patients, 766 (54.9%) experienced mucositis with a continuous infusion of FU, compared to 532 (34.4%) out of 1545 patients receiving chronotherapy. Among patients in the treatment group that received constant administration of FU, the ratio of those who developed mucositis to those who did not (766/629 = 1.2178) was significantly higher than the same ratio obtained in the group following chronomodulated administration of FU (532/1013 = 0.5252) (X2 = 123.83, p < 0.0001). These findings indicate that the published research evidence supports chronomodulated FU therapy as leading to significantly fewer patients to develop mucositis compared to the continuous infusion regimen of FU.

This systematic evaluation of the literature confirms that research toward understanding and controlling the biochemical and molecular variables that direct and regulate the circadian pattern of FU catabolism by DPD requires immediate attention. Specifically, it has been proposed that future research focus on easy-to-obtain markers of specific rhythms to individualize the chronomodulated FU delivery (Milano and Chamorey 2002).

Super Serenity Sleepers

Super Serenity Sleepers

Do You Have Problem Getting A Good Night Sleep? Learn To Sleep Like A Cat At Night And Run Like A Lion When You Wake Up.

Get My Free Ebook

Post a comment