- Email: [email protected]
Evidence for various calcium sites in human hair shaft revealed by sub-micrometer X-ray fluorescence
Biochimica et Biophysica Acta 1619 (2003) 53 – 58 www.bba-direct.com
Evidence for various calcium sites in human hair shaft revealed by sub-micrometer X-ray fluorescence C. Me´rigoux a, F. Briki a, F. Sarrot-Reynauld b, M. Salome´ c, B. Fayard c, J. Susini c, J. Doucet a,* b
a LURE, Universite´ Paris-Sud, Baˆt 209-D, B.P. 34, F-91898 Orsay cedex, France Service de Me´decine Interne, CHU Michallon, B.P. 217, F-38043 Grenoble cedex 9, France c ESRF, B.P. 220, F-38043 Grenoble cedex, France
Received 25 April 2002; received in revised form 5 September 2002; accepted 23 September 2002
Abstract New information about calcium status in human scalp hair shaft, deduced from X-ray micro-fluorescence imaging, including its distribution over the hair section, the existence of one or several binding-types and its variation between people, is presented. The existence of two different calcium types is inferred. The first one corresponds to atoms (or ions) easily removable by hydrochloric acid, located in the cortex (granules), in the cuticle zone and also in the core of the medulla, which can reasonably be identified as calcium soaps. The second type consists of non-easily removable calcium atoms (or ions) that are located in the medulla wall, probably also in the cuticle, and rather uniformly in the cortex; these calcium atoms may be involved in Ca2 +-binding proteins, and their concentration is fairly constant from one subject to another. In addition to its nonuniform distribution across the hair section, the second striking feature of the first type calcium content is its high variability from one subject to another, by up to a factor 10. We expect this information to be useful for analyzing in more detail the relationship between hair calcium and environmental and medical factors. D 2002 Published by Elsevier Science B.V. Keywords: Hair; Lipid; Calcium; Calcium soap; Calcium binding protein; X-ray fluorescence
1. Introduction Hair is known to contain usually about 30 trace elements in the range 0.1 to 1000 ppm [1] with a total concentration in the range 2500 to 10 000 ppm [2,3]. Some of these elements are of endogenous origin, others of exogenous origin, reflecting the high propensity of hair for absorbing chemical elements. Because of its growth process, hair might in addition reveal the biomedical and environment history of the subject. A major interest in calcium, one of the most abundant trace elements in hair with chlorine, lies in its usual implication in biological processes. Its overall concentration is generally of the order of several hundreds ppm [1]. Twentyfive years ago, the calcium content in hair was reported to be age- and medical status-dependent. A similar conclusion was recently reached from a large population analysis as regards
* Corresponding author. Tel./fax: +33-1-64-46-88-20. E-mail address: [email protected] (J. Doucet). 0304-4165/02/$ - see front matter D 2002 Published by Elsevier Science B.V. PII: S 0 3 0 4 - 4 1 6 5 ( 0 2 ) 0 0 4 4 1 - 5
prevalence to heart disease and environmental conditions as water hardness or sunshine hours [4]. Hair calcium content in human hair shaft could therefore be considered as a marker for different factors, however, a better understanding of its chemical and physical statuses is required before envisaging screening methods on a single hair. Trace elements in human and animal hair have been studied for a long time using various analytical methods [5]. Recently, X-ray fluorescence analysis technique was used for quantitative estimation of elemental hair content [6]. This study aimed to build a database for elements concentration within a given population. Using the proton microprobe technique, various radial concentration profiles [7– 12] with a few micrometers resolution, and a mapping of pig hair section with a 2-Am resolution [13], showed a nonuniform distribution of many elements, including calcium. The profiles show a clear reinforcement of Ca concentration in the outer part of the shaft and, for a few samples, another sharp reinforcement in the core of the shaft (about 5 – 10-Am wide). We surmise that these data reveal the existence of various calcium sites probably linked to different hair sub-
C. Me´rigoux et al. / Biochimica et Biophysica Acta 1619 (2003) 53–58
54
components. Until now the spatial resolution afforded by the techniques used was too low to gain information about fine structural details in hair at 1 Am or at the sub-micrometer scale. During the last 10 years, the presence of several calcium binding proteins was detected in human hair follicles: calmodulin [14], caldesmon [15], calcyclin mRNA in the postmitotic keratogenous region of the hair follicle [16], S-100 within the neuronal network surrounding the follicle, and annexin-1 in hair follicle of adult skin [17]. Only one observation has been made for hair shaft: an S100A3 family protein was visualized in the cuticle [18]. The other major component of hair consists of lipids [19 –21]. Evidence for the existence of ‘‘internal’’ lipids, in contrast to the ‘‘external’’ sebum which is located on the hair surface, was provided twenty five years ago. Part of the external lipids were identified as fatty acids bound to free carboxylic groups of the proteins by Mg2 + or Ca2 + bridges [22]; however, no link between internal lipids and calcium was established so far. We aim here to provide new information about calcium status in human hair shaft, including its distribution over the hair section, the existence of one or several binding types and its variation between people. This information is likely to be useful in understanding the relationship of hair calcium concentration and environmental and medical factors. Precise information about calcium could also be helpful for cosmetic applications. It is worth mentioning that our study was made possible, thanks to the new experimental synchrotron radiation facilities overcoming the limitations of classical home equipment for micro-fluorescence analysis.
2. Materials and methods 2.1. Hair preparation Thin slices (about 50-Am thick) of human scalp hair shaft cut perpendicular to the hair axis were prepared. Hair
samples were collected from 10 healthy Caucasian males or females subjects. They were inserted into a plastic pipe and embedded in a Micro-Lac solution (SPI-Chem). Cross sections were cut at a thickness of about 50 Am using a fiber microtome (SPI-Supplies). The blades were cleaned between each pass to prevent contamination on the section. One to five cross sections were analyzed for each hair. The slices were mounted between two 10-Am-thick ‘Ultralene’ foils (SPEX Certiprep Inc.); a plastic foil devoid of inorganic contaminants. This was then fixed on the sample holder. For the lipid removal treatment, hair was immersed in 0.1 M hydrochloric acid solution for 1 h, and then rinsed by multiple changes of distilled water over 1 h. Hydrochloric acid was expected to hydrolyze hair lipids as it does for lipids in gastric fluid. 2.2. Micro-fluorescence set-up Samples were examined using the X-ray microscopy beamline, ID21, at the European Synchrotron Radiation Facility (ESRF; Grenoble, France) [23]. The Scanning Xray microscope used Fresnel zone plates as focusing optics, which demagnify the synchrotron X-ray source to generate a sub-micron probe [24 25]. The microprobe sizes were 0.3 0.3 Am2 at the Calcium K-edge. As shown in Fig. 1, fluorescence and transmission signals can be used simultaneously to investigate the sample, which is mounted on a piezo-electric stage and raster scanned in the beam to acquire a two-dimensional image point by point. To minimize the contribution of the unwanted Thomson elastic scattering, the normal to the sample stage is tilted about 15j with respect to the beam direction, and the energydispersive high-purity Germanium single element detector mounted in the horizontal plane perpendicular to the beam is used to collect the fluorescence signal. The fluorescence signal was processed via a multiple channel analyzer allowing fluorescence emitted by various elements to be recorded simultaneously. The input X-ray energy, slightly larger than
Fig. 1. Optical scheme of the ID21 branch-line housing the scanning X-ray microscope (SXM) used in fluorescence/transmission modes at the ESRF.
C. Me´rigoux et al. / Biochimica et Biophysica Acta 1619 (2003) 53–58
the Ca-K absorption edge, was tuned using a fixed-exit silicon monochromator. Finally, a fixed-exit double mirror system, located upstream from the monochromator and acting as a low band pass filter, allowed harmonic rejection greater than 10 3 with total transmission greater than 70%. The measurements were carried out in vacuum (10 5 mbar) to minimize absorption by air of the sulfur XRF signal. 2.3. Fluorescence data analysis Relative calcium/sulfur mass concentrations (CCa/CS) were calculated from the corresponding fluorescence intensity ratios, averaged over chosen areas of the image (ICa and IS), the size and position of which were adapted to the analyzed zones. The sulfur concentration being nearly constant from one hair to another (about 50 000 ppm [13]), this method enables us to compare the calcium concentrations from different samples by normalizing the sulfur concentrations. Sulfur concentration was estimated by averaging over a large area in the cortex. All the classical correcting factors were applied to the intensities: the Ka emission probabilities, the mass photoelectric absorption coefficients, the mass absorption coefficients of samples, the thickness of the sample, its orientation versus the direct beam and the detector, the absorption by the ‘Ultralene’ foil, the absorption by the detector window and the efficiencies of the detector at the two Ka energies. The mass density of the hair was taken as being constant for all samples (1.32 g/cm3 [26]). The various corrections lead to the following correcting factor: CCa =CS ¼ 0:033ICa =IS The relative error of CCa/CS for various zones in a given sample is smaller than 5%. From one sample to another, due to possible errors coming from various experimental parameters (sample thickness, orientation of the sample, etc.), the relative error is much larger, of the order of 30%. However, as shown below, this error is smaller than the variations of calcium content from one hair to another.
3. Results and interpretation 3.1. Calcium mapping Typical calcium micro-fluorescence mappings at the submicrometer resolution of calcium content on a hair section are shown in Fig. 2a– c. Three regions are visible proceeding radially inwards which correspond to the three main structural parts of hair shaft [27]: – within the multilayered cuticle of overall thickness 5 Am, occasional short arc-like high concentrations of calcium, – within the cortex, numerous calcium-rich zones, each a few micrometers in diameter,
55
– within the central medulla of 10-Am diameter, an even distribution of high calcium content across the whole region. The corresponding maps for sulfur distribution (Fig. 2d– f) show this element is fairly evenly distributed across the whole fiber apart from a slight reinforcement in the cuticle, a depletion in the medulla and slight depletions in each calcium-rich granule. The low sulfur content of the granules, together with their size, shape and distribution, might suggest that the granules are associated to the nuclear remnants, which are spindle-shaped structures about 40Am long and 1 Am in transverse section. This hypothesis is supported by the reinforcement of the calcium concentration in the cuticle where many nuclear remnants are known to be located, but it still has to be confirmed and we have no explanation as to why calcium should be found in these structures. Zones of high calcium content in the cortex and medulla were highly variable in their visibility between hair samples. On the other hand, regions of high calcium concentration within the cuticle were in evidence in all samples. 3.2. Calcium concentration The relative calcium/sulfur mass concentration ratios (CCa/CS) in the various zones of hair have been calculated from the intensity ratios of the calcium to sulfur Ka fluorescence intensities as described in Materials and methods. In Table 1 are given the CCa/CS values determined in the three zones for hair of different origins. The ratios are highly variable between hairs. In the cortex, the average CCa/CS ratios range from 1.78% down to 0.13%; the average value is 0.74% and the standard deviation is 0.60%. However, the concentration is not uniform, the concentration in granules (when present) is about twice that outside granules. The CCa/CS ratios in the cuticle zone range from 3.6% down to 0.33% (average value 1.56%, standard deviation 1.03%); it is always higher than in cortex. The CCa/CS ratios corresponding to the medulla are even more variable: they range from 6.33% down to 0.47% with an average value of 2.33% and standard deviation 2.25%; no correlation with the values in the cortex or cuticle zones can be established. Although the uncertainties on these values is rather large (30%) and the sulfur content slightly variable, which makes the accurate determination of absolute calcium concentration quite difficult, the large amplitude variations of calcium content within a given hair and from one hair to another are quite significant. This conclusion is supported by the observation of similar calcium distribution features across the hair section. It is pertinent to note that our CCa/CS determinations are in agreement with the estimations of overall calcium content in hair found in literature (1% to 2%) [1,13]. It seems reasonable to suppose that these well-defined calcium locations within hair section reveal the existence of different calcium-binding types. In order to assess this hypothesis,
56
C. Me´rigoux et al. / Biochimica et Biophysica Acta 1619 (2003) 53–58
Fig. 2. Microfluorescence images of a section of human scalp hair collected at the Ca Ka and S Ka energies. (a) gives the complete calcium distribution across the section obtained with a 1 1-Am2 pixel size; (b) and (c) give detailed mappings (indicated by rectangles in (a)) of the cuticle zone and of the medulla obtained with a 0.33 0.33-Am2 pixel size (beam spot size, 0.3 0.3 m2). (d), (e) and (f) give the corresponding sulfur mappings. Data collection time was 1 s/ pixel. Calcium content is clearly higher in the medulla, in small granules within the cortex and in the cuticle zone.
complementary fluorescence analyses were carried out on hydrochloric acid-treated samples. 3.3. Treatment by hydrochloric acid We have also examined the effect of hydrochloric acid treatment in that it might lead to some differential removal of calcium according to the nature of its binding in the hair. Micro-fluorescence calcium and sulfur mappings were carried out on two samples, sample 1 which is typical of a high
calcium content and sample 10 which is typical of a low calcium content. The corresponding calcium concentrations are given in Table 1; note that it is not unreasonable to assume the sulfur content will not have been affected by the hydrochloric acid treatment. As expected, the overall calcium concentration is lower in treated samples, but the interesting point is that the decrease is not uniform, as clearly evidenced in Fig. 3. For the hair with a high calcium content, the decrease is high in the cortex (factor of 10 in average in the cortex and
C. Me´rigoux et al. / Biochimica et Biophysica Acta 1619 (2003) 53–58 Table 1 Calcium/sulfur mass concentration ratios in various parts of the hair: cortex, cuticle and medulla (when present) Sample
Cortex Cortex Cuticle Cuticle Medulla Medulla (average) (average) zone zone after after HCl after HCl HCl
1 2 3 4 5 6 7 8 9 10 Average Standard deviation
1.78 0.43 0.13 0.17 0.40 0.70 1.40 1.63 0.17 0.55 0.74 0.60
0.17
0.40 0.28
2.43 0.43 0.33 1.50 0.63 2.43 3.60 2.27 1.16 0.83 1.56 1.03
0.21
6.33 1.00
1.11
0.53 0.47 3.30
0.41 0.31
2.33 2.25
Calcium/sulfur mass concentration ratios are expressed in percent. For two samples (a high-calcium content one and a low-calcium content one), the corresponding values after treatment in an hydrochloric acid solution are given. The average and standard deviation of each column are given in the last two lines.
20 for the granules) and in the cuticle (factor of 10), whilst it is smaller in the medulla and up to 10 Am outside the medulla wall (factor 5). Let us note the fact that the wall of the medulla is less affected by the hydrochloric acid treatment than its core; it becomes clearly visible after hydrochloric acid treatment. On the contrary, the treatment for a low calcium content hair only induced a small change in the cortex (factor of 0.7) and in the cuticle zone (factor of 2). Interestingly, the after-treatment calcium concentrations in the cortex and in the cuticle zone of high and low calcium content hair are close (around 0.3%). The existence of two different calcium types can therefore be inferred from these observations. The first one corresponds to atoms easily removable by hydrochloric
57
acid, located in the cortex granules, in the cuticle zone and also in the core of the medulla. The second type consists of less easily removable calcium located in the medulla wall and rather uniformly in the cortex. The concentration of the first type of calcium is highly variable whilst that of the second-type calcium seems to be independent of hair sample (0.3%).
4. Discussion The chemical identification of these two types of calcium sites requires extra information. Concerning the first type of calcium, studies by infrared spectromicroscopy of hair sections [28 – 30] have shown the presence of lipids within the medulla. As regards the granules in the cortex, the existence of crystalline lipids domains in hair has already been proposed by Fraser et al. [31] 40 years ago. These lipids give rise to well-defined sharp diffraction rings, which proves their crystalline nature. X-ray diffraction analyses have shown that their concentration is generally higher in the 5- to 10-Am-thick outer layer than in the cortex [32]. These observations could be interpreted on the basis of hair lipids in the form of calcium soaps, i.e. where the calcium is in a form readily removable by hydrochloric acid. It also follows that without such calcium sequestration the lipid component will be incapable of extraction with an organic solvent. Thus, despite the large number of analyses carried out over the years of lipid extracted from hair by solvents, the soaps may have escaped detection because the calcium had not been first removed. The existence of calcium soaps at the surface of hair was in fact inferred a long time ago from chemical investigations [22]; our observations strongly support the existence of internal calcium soaps in hair. We have not been able to secure information as to the binding in the hair of second type of calcium, i.e. that which
Fig. 3. Microfluorescence images of a section of human scalp hair (high calcium content hair, same subject as for Fig. 2) treated in a 0.1 N hydrochloric acid solution, collected at the Ca Ka and S Ka energies. (a) and (b) give respectively the calcium and sulfur distributions across the section obtained with a 1 1Am2 pixel size (beam spot size, 0.3 0.3 m2). Data collection was 1 s/pixel. After lipid removal by HCl, the remaining calcium atoms are mainly located in and around the wall of the medulla. Note the tear in the sample.
58
C. Me´rigoux et al. / Biochimica et Biophysica Acta 1619 (2003) 53–58
resists removal by hydrochloric acid. However, it is reasonable to assume that they are associated to proteins, which would make them less removable by hydrochloric acid. Calcium ions are in fact involved in many biochemical processes. The existence of such proteins has already been established for hair follicle but not yet for hair shaft, except for an S100 family protein in the cuticle. It is worth mentioning that a link between Ca2 + binding proteins (annexins) and lipids in the sarcolemma of smooth muscle cells was reported: the Ca2 + ions regulate the association of lipid microdomains in the plasma membrane [33]. A similar link could be envisaged for hair, at least in the hair follicle. The origin of the calcium-binding proteins is certainly endogenous. On the contrary, the origin of the calcium atoms bound to lipids can be either endogenous or exogenous [34], even if they are located inside the hair fiber. The high variability of the internal calcium soap concentration from one subject to another does not bring any light on this question. The origin of such huge variability remains unclear; further studies are needed to investigate this point which could reflect relevant physiological, biological, hair care, pathological or environmental factors.
Acknowledgements We thank Dr. Pierre Chevallier for support in the fluorescence data analysis and the reviewer who suggested the association of calcium-rich granules with nuclear remnants and made valuable suggestions for the discussion about calcium soap extraction.
References [1] [2] [3] [4] [5]
S.B. M’Baku, R.M. Parr, J. Radioanal. Chem. 69 (1982) 171. T.F. Dutcher, S.J. Rothman, J. Invest. Dermatol. 17 (1951) 65. F.G.E. Pautard, Nature 199 (1963) 531. A. MacPherson, J. Basco, Sci. Total Environ. 1 – 3 (2000) 11 – 19. V. Valkovic, I. Fundamentals and Methods for Measurement of Elemental Composition, CRC Press, Boca Raton, FL, 1988, p. 1. [6] B. Stocklassa, M. Aransay-Vitores, G. Nilsson, C. Karlsson, D. Wiegleb, B. Forslind, J. Cosmet. Sci. 52 (2001) 297 – 311. [7] A.J.J. Bos, C.C.A.H.V.D. Stap, W.J.M. Lenglet, R.D. Vis, V. Valkovic, IEEE Trans. Nucl. Sci. 30 (1983) 1249. [8] A.J.J. Bos, R.D. Vis, H. Verheul, M. Prins, S.T. Davies, D.K. Bowen, J. Makjanic, V. Valkovic, Nucl. Instrum. Methods Phys. Res., B 3 (1984) 232.
[9] A.J.J. Bos, R.D. Vis, W.J.M. Lenglet, H. Verheul, V. Valkovic, J. Makjanic, in: P Bra¨tter, P. Shramel (Eds.), Proceedings 2nd Workshop in Trace Element Analytical Chemistry in Medicine and Biology, Walter de Gruyter, Berlin, 1983, p. 787. [10] K.G.K.R. Malmquist, B. Forslind, Nucl. Instrum. Methods 181 (1981) 465. [11] M. Uo, F. Watari, A. Yokoyama, H. Matsuno, T. Kawasaki, Biomaterials 20 (1999) 747 – 755. [12] J.A. Cookson, F.D. Pilling, Phys. Med. Biol. 20 (1975) 1015 – 1020. [13] I. Bogdanovic, M. Jaksic, S. Fazinic, J.W. Hopewell, J.H. Wilkinson, V. Valkovic, Nucl. Instrum. Methods Phys. Res., B 104 (1995) 339 – 343. [14] U. Wollina, R. Klinger, R. Wetzker, R. Reissmann, B. Knopf, Arch. Dermatol. Res. 280 (1989) 497 – 498. [15] F. Matsumura, S. Yamashiro, Curr. Opin. Cell Biol. 5 (1993) 70 – 76. [16] L. Wood, D. Carter, M. Mills, N. Hatzenbulher, G. Vogel, J. Invest. Dermatol. 96 (1991) 383 – 387. [17] R.A. Fava, L.B. Nanney, D. Wilson, L.E. King Jr., J. Invest. Dermatol. 101 (1993) 732 – 737. [18] K. Kizawa, H. Uchiwa, U. Murakami, Biochim. Biophys. Acta 1312 (1996) 94 – 98. [19] A.W. Weitkamp, A.M. Smiljanic, S. Rothman, J. Am. Oil Chem. Soc. 69 (1947) 1936. [20] N. Nicolaides, S. Rothman, J. Invest. Dermatol. 21 (1953) 9 – 12. [21] O. Sakamoto, Y. Fujinuma, T. Ozawa, J. Am. Oil Chem. Soc. 54 (1977) 143. [22] K.V. Curry, S. Golding, J. Soc. Cosmet. Chem. 22 (1971) 681 – 699. [23] J. Susini, R. Barrett, B. Kaulich, S. Oestreich, M. Salome´, in: W. Meyer, A. Warwick, D.T. Attwood (Eds.), X-Ray Microscopy Proceedings of the Sixth International Conference, vol. 507, American Institute of Physics, New York, 2000, pp. 19 – 26. [24] E. Di Fabrizio, F. Romanato, M. Gentili, S. Cabrini, B. Kaulich, J. Susini, R. Barrett, Nature 401 (1999) 895 – 898. [25] C. David, B. Kaulich, R. Barrett, M. Salome´, J. Susini, Appl. Phys. Lett. 77 (2000) 3851 – 3853. [26] N. Abbott, A. Goodings, J. Text. Inst. 40T (1949) 232. [27] R.D.B. Fraser, T.P. MacRae, G.E. Rogers, KERATINS: Their Composition, Structure and Biosynthesis, Charles Thomas, Springfield, IL, 1972. [28] E. Bramanti, F. Ronca, L. Teodori, M.L. Truica, F. Papineschi, G. Spremolla II, P. Vergamini, E.A. Benedetti, J. Soc. Cosmet. Chem. 43 (1992) 285. [29] J.L. Bantignies, G. Fuchs, G.L. Carr, G.P. Williams, D. Lutz, S. Marull, Int. J. Cosmet. Sci. 20 (1998) 381 – 394. [30] L. Kreplak, C. Merigoux, F. Briki, D. Flot, J. Doucet, Biochim. Biophys. Acta 1547 (2001) 268 – 274. [31] R.D.B. Fraser, T.P. Macrae, G.E. Rogers, B.K. Filshie, J. Mol. Biol. 7 (1963) 90 – 91. [32] B. Busson, P. Engstro¨m, J. Doucet, J. Synchrotron Radiat. 6 (1999) 1021 – 1030. [33] E.B. Babiychuck, A. Draeger, J. Cell Biol. 150 (2000) 1113 – 1123. [34] J. Basco, L. Sarkadi, E. Koltay, Int. J. Appl. Radiat. Isot. 33 (1982) 5.
Evidence for various calcium sites in human hair shaft revealed by sub-micrometer X-ray fluorescence C. Me´rigoux a, F. Briki a, F. Sarrot-Reynauld b, M. Salome´ c, B. Fayard c, J. Susini c, J. Doucet a,* b
a LURE, Universite´ Paris-Sud, Baˆt 209-D, B.P. 34, F-91898 Orsay cedex, France Service de Me´decine Interne, CHU Michallon, B.P. 217, F-38043 Grenoble cedex 9, France c ESRF, B.P. 220, F-38043 Grenoble cedex, France
Received 25 April 2002; received in revised form 5 September 2002; accepted 23 September 2002
Abstract New information about calcium status in human scalp hair shaft, deduced from X-ray micro-fluorescence imaging, including its distribution over the hair section, the existence of one or several binding-types and its variation between people, is presented. The existence of two different calcium types is inferred. The first one corresponds to atoms (or ions) easily removable by hydrochloric acid, located in the cortex (granules), in the cuticle zone and also in the core of the medulla, which can reasonably be identified as calcium soaps. The second type consists of non-easily removable calcium atoms (or ions) that are located in the medulla wall, probably also in the cuticle, and rather uniformly in the cortex; these calcium atoms may be involved in Ca2 +-binding proteins, and their concentration is fairly constant from one subject to another. In addition to its nonuniform distribution across the hair section, the second striking feature of the first type calcium content is its high variability from one subject to another, by up to a factor 10. We expect this information to be useful for analyzing in more detail the relationship between hair calcium and environmental and medical factors. D 2002 Published by Elsevier Science B.V. Keywords: Hair; Lipid; Calcium; Calcium soap; Calcium binding protein; X-ray fluorescence
1. Introduction Hair is known to contain usually about 30 trace elements in the range 0.1 to 1000 ppm [1] with a total concentration in the range 2500 to 10 000 ppm [2,3]. Some of these elements are of endogenous origin, others of exogenous origin, reflecting the high propensity of hair for absorbing chemical elements. Because of its growth process, hair might in addition reveal the biomedical and environment history of the subject. A major interest in calcium, one of the most abundant trace elements in hair with chlorine, lies in its usual implication in biological processes. Its overall concentration is generally of the order of several hundreds ppm [1]. Twentyfive years ago, the calcium content in hair was reported to be age- and medical status-dependent. A similar conclusion was recently reached from a large population analysis as regards
* Corresponding author. Tel./fax: +33-1-64-46-88-20. E-mail address: [email protected] (J. Doucet). 0304-4165/02/$ - see front matter D 2002 Published by Elsevier Science B.V. PII: S 0 3 0 4 - 4 1 6 5 ( 0 2 ) 0 0 4 4 1 - 5
prevalence to heart disease and environmental conditions as water hardness or sunshine hours [4]. Hair calcium content in human hair shaft could therefore be considered as a marker for different factors, however, a better understanding of its chemical and physical statuses is required before envisaging screening methods on a single hair. Trace elements in human and animal hair have been studied for a long time using various analytical methods [5]. Recently, X-ray fluorescence analysis technique was used for quantitative estimation of elemental hair content [6]. This study aimed to build a database for elements concentration within a given population. Using the proton microprobe technique, various radial concentration profiles [7– 12] with a few micrometers resolution, and a mapping of pig hair section with a 2-Am resolution [13], showed a nonuniform distribution of many elements, including calcium. The profiles show a clear reinforcement of Ca concentration in the outer part of the shaft and, for a few samples, another sharp reinforcement in the core of the shaft (about 5 – 10-Am wide). We surmise that these data reveal the existence of various calcium sites probably linked to different hair sub-
C. Me´rigoux et al. / Biochimica et Biophysica Acta 1619 (2003) 53–58
54
components. Until now the spatial resolution afforded by the techniques used was too low to gain information about fine structural details in hair at 1 Am or at the sub-micrometer scale. During the last 10 years, the presence of several calcium binding proteins was detected in human hair follicles: calmodulin [14], caldesmon [15], calcyclin mRNA in the postmitotic keratogenous region of the hair follicle [16], S-100 within the neuronal network surrounding the follicle, and annexin-1 in hair follicle of adult skin [17]. Only one observation has been made for hair shaft: an S100A3 family protein was visualized in the cuticle [18]. The other major component of hair consists of lipids [19 –21]. Evidence for the existence of ‘‘internal’’ lipids, in contrast to the ‘‘external’’ sebum which is located on the hair surface, was provided twenty five years ago. Part of the external lipids were identified as fatty acids bound to free carboxylic groups of the proteins by Mg2 + or Ca2 + bridges [22]; however, no link between internal lipids and calcium was established so far. We aim here to provide new information about calcium status in human hair shaft, including its distribution over the hair section, the existence of one or several binding types and its variation between people. This information is likely to be useful in understanding the relationship of hair calcium concentration and environmental and medical factors. Precise information about calcium could also be helpful for cosmetic applications. It is worth mentioning that our study was made possible, thanks to the new experimental synchrotron radiation facilities overcoming the limitations of classical home equipment for micro-fluorescence analysis.
2. Materials and methods 2.1. Hair preparation Thin slices (about 50-Am thick) of human scalp hair shaft cut perpendicular to the hair axis were prepared. Hair
samples were collected from 10 healthy Caucasian males or females subjects. They were inserted into a plastic pipe and embedded in a Micro-Lac solution (SPI-Chem). Cross sections were cut at a thickness of about 50 Am using a fiber microtome (SPI-Supplies). The blades were cleaned between each pass to prevent contamination on the section. One to five cross sections were analyzed for each hair. The slices were mounted between two 10-Am-thick ‘Ultralene’ foils (SPEX Certiprep Inc.); a plastic foil devoid of inorganic contaminants. This was then fixed on the sample holder. For the lipid removal treatment, hair was immersed in 0.1 M hydrochloric acid solution for 1 h, and then rinsed by multiple changes of distilled water over 1 h. Hydrochloric acid was expected to hydrolyze hair lipids as it does for lipids in gastric fluid. 2.2. Micro-fluorescence set-up Samples were examined using the X-ray microscopy beamline, ID21, at the European Synchrotron Radiation Facility (ESRF; Grenoble, France) [23]. The Scanning Xray microscope used Fresnel zone plates as focusing optics, which demagnify the synchrotron X-ray source to generate a sub-micron probe [24 25]. The microprobe sizes were 0.3 0.3 Am2 at the Calcium K-edge. As shown in Fig. 1, fluorescence and transmission signals can be used simultaneously to investigate the sample, which is mounted on a piezo-electric stage and raster scanned in the beam to acquire a two-dimensional image point by point. To minimize the contribution of the unwanted Thomson elastic scattering, the normal to the sample stage is tilted about 15j with respect to the beam direction, and the energydispersive high-purity Germanium single element detector mounted in the horizontal plane perpendicular to the beam is used to collect the fluorescence signal. The fluorescence signal was processed via a multiple channel analyzer allowing fluorescence emitted by various elements to be recorded simultaneously. The input X-ray energy, slightly larger than
Fig. 1. Optical scheme of the ID21 branch-line housing the scanning X-ray microscope (SXM) used in fluorescence/transmission modes at the ESRF.
C. Me´rigoux et al. / Biochimica et Biophysica Acta 1619 (2003) 53–58
the Ca-K absorption edge, was tuned using a fixed-exit silicon monochromator. Finally, a fixed-exit double mirror system, located upstream from the monochromator and acting as a low band pass filter, allowed harmonic rejection greater than 10 3 with total transmission greater than 70%. The measurements were carried out in vacuum (10 5 mbar) to minimize absorption by air of the sulfur XRF signal. 2.3. Fluorescence data analysis Relative calcium/sulfur mass concentrations (CCa/CS) were calculated from the corresponding fluorescence intensity ratios, averaged over chosen areas of the image (ICa and IS), the size and position of which were adapted to the analyzed zones. The sulfur concentration being nearly constant from one hair to another (about 50 000 ppm [13]), this method enables us to compare the calcium concentrations from different samples by normalizing the sulfur concentrations. Sulfur concentration was estimated by averaging over a large area in the cortex. All the classical correcting factors were applied to the intensities: the Ka emission probabilities, the mass photoelectric absorption coefficients, the mass absorption coefficients of samples, the thickness of the sample, its orientation versus the direct beam and the detector, the absorption by the ‘Ultralene’ foil, the absorption by the detector window and the efficiencies of the detector at the two Ka energies. The mass density of the hair was taken as being constant for all samples (1.32 g/cm3 [26]). The various corrections lead to the following correcting factor: CCa =CS ¼ 0:033ICa =IS The relative error of CCa/CS for various zones in a given sample is smaller than 5%. From one sample to another, due to possible errors coming from various experimental parameters (sample thickness, orientation of the sample, etc.), the relative error is much larger, of the order of 30%. However, as shown below, this error is smaller than the variations of calcium content from one hair to another.
3. Results and interpretation 3.1. Calcium mapping Typical calcium micro-fluorescence mappings at the submicrometer resolution of calcium content on a hair section are shown in Fig. 2a– c. Three regions are visible proceeding radially inwards which correspond to the three main structural parts of hair shaft [27]: – within the multilayered cuticle of overall thickness 5 Am, occasional short arc-like high concentrations of calcium, – within the cortex, numerous calcium-rich zones, each a few micrometers in diameter,
55
– within the central medulla of 10-Am diameter, an even distribution of high calcium content across the whole region. The corresponding maps for sulfur distribution (Fig. 2d– f) show this element is fairly evenly distributed across the whole fiber apart from a slight reinforcement in the cuticle, a depletion in the medulla and slight depletions in each calcium-rich granule. The low sulfur content of the granules, together with their size, shape and distribution, might suggest that the granules are associated to the nuclear remnants, which are spindle-shaped structures about 40Am long and 1 Am in transverse section. This hypothesis is supported by the reinforcement of the calcium concentration in the cuticle where many nuclear remnants are known to be located, but it still has to be confirmed and we have no explanation as to why calcium should be found in these structures. Zones of high calcium content in the cortex and medulla were highly variable in their visibility between hair samples. On the other hand, regions of high calcium concentration within the cuticle were in evidence in all samples. 3.2. Calcium concentration The relative calcium/sulfur mass concentration ratios (CCa/CS) in the various zones of hair have been calculated from the intensity ratios of the calcium to sulfur Ka fluorescence intensities as described in Materials and methods. In Table 1 are given the CCa/CS values determined in the three zones for hair of different origins. The ratios are highly variable between hairs. In the cortex, the average CCa/CS ratios range from 1.78% down to 0.13%; the average value is 0.74% and the standard deviation is 0.60%. However, the concentration is not uniform, the concentration in granules (when present) is about twice that outside granules. The CCa/CS ratios in the cuticle zone range from 3.6% down to 0.33% (average value 1.56%, standard deviation 1.03%); it is always higher than in cortex. The CCa/CS ratios corresponding to the medulla are even more variable: they range from 6.33% down to 0.47% with an average value of 2.33% and standard deviation 2.25%; no correlation with the values in the cortex or cuticle zones can be established. Although the uncertainties on these values is rather large (30%) and the sulfur content slightly variable, which makes the accurate determination of absolute calcium concentration quite difficult, the large amplitude variations of calcium content within a given hair and from one hair to another are quite significant. This conclusion is supported by the observation of similar calcium distribution features across the hair section. It is pertinent to note that our CCa/CS determinations are in agreement with the estimations of overall calcium content in hair found in literature (1% to 2%) [1,13]. It seems reasonable to suppose that these well-defined calcium locations within hair section reveal the existence of different calcium-binding types. In order to assess this hypothesis,
56
C. Me´rigoux et al. / Biochimica et Biophysica Acta 1619 (2003) 53–58
Fig. 2. Microfluorescence images of a section of human scalp hair collected at the Ca Ka and S Ka energies. (a) gives the complete calcium distribution across the section obtained with a 1 1-Am2 pixel size; (b) and (c) give detailed mappings (indicated by rectangles in (a)) of the cuticle zone and of the medulla obtained with a 0.33 0.33-Am2 pixel size (beam spot size, 0.3 0.3 m2). (d), (e) and (f) give the corresponding sulfur mappings. Data collection time was 1 s/ pixel. Calcium content is clearly higher in the medulla, in small granules within the cortex and in the cuticle zone.
complementary fluorescence analyses were carried out on hydrochloric acid-treated samples. 3.3. Treatment by hydrochloric acid We have also examined the effect of hydrochloric acid treatment in that it might lead to some differential removal of calcium according to the nature of its binding in the hair. Micro-fluorescence calcium and sulfur mappings were carried out on two samples, sample 1 which is typical of a high
calcium content and sample 10 which is typical of a low calcium content. The corresponding calcium concentrations are given in Table 1; note that it is not unreasonable to assume the sulfur content will not have been affected by the hydrochloric acid treatment. As expected, the overall calcium concentration is lower in treated samples, but the interesting point is that the decrease is not uniform, as clearly evidenced in Fig. 3. For the hair with a high calcium content, the decrease is high in the cortex (factor of 10 in average in the cortex and
C. Me´rigoux et al. / Biochimica et Biophysica Acta 1619 (2003) 53–58 Table 1 Calcium/sulfur mass concentration ratios in various parts of the hair: cortex, cuticle and medulla (when present) Sample
Cortex Cortex Cuticle Cuticle Medulla Medulla (average) (average) zone zone after after HCl after HCl HCl
1 2 3 4 5 6 7 8 9 10 Average Standard deviation
1.78 0.43 0.13 0.17 0.40 0.70 1.40 1.63 0.17 0.55 0.74 0.60
0.17
0.40 0.28
2.43 0.43 0.33 1.50 0.63 2.43 3.60 2.27 1.16 0.83 1.56 1.03
0.21
6.33 1.00
1.11
0.53 0.47 3.30
0.41 0.31
2.33 2.25
Calcium/sulfur mass concentration ratios are expressed in percent. For two samples (a high-calcium content one and a low-calcium content one), the corresponding values after treatment in an hydrochloric acid solution are given. The average and standard deviation of each column are given in the last two lines.
20 for the granules) and in the cuticle (factor of 10), whilst it is smaller in the medulla and up to 10 Am outside the medulla wall (factor 5). Let us note the fact that the wall of the medulla is less affected by the hydrochloric acid treatment than its core; it becomes clearly visible after hydrochloric acid treatment. On the contrary, the treatment for a low calcium content hair only induced a small change in the cortex (factor of 0.7) and in the cuticle zone (factor of 2). Interestingly, the after-treatment calcium concentrations in the cortex and in the cuticle zone of high and low calcium content hair are close (around 0.3%). The existence of two different calcium types can therefore be inferred from these observations. The first one corresponds to atoms easily removable by hydrochloric
57
acid, located in the cortex granules, in the cuticle zone and also in the core of the medulla. The second type consists of less easily removable calcium located in the medulla wall and rather uniformly in the cortex. The concentration of the first type of calcium is highly variable whilst that of the second-type calcium seems to be independent of hair sample (0.3%).
4. Discussion The chemical identification of these two types of calcium sites requires extra information. Concerning the first type of calcium, studies by infrared spectromicroscopy of hair sections [28 – 30] have shown the presence of lipids within the medulla. As regards the granules in the cortex, the existence of crystalline lipids domains in hair has already been proposed by Fraser et al. [31] 40 years ago. These lipids give rise to well-defined sharp diffraction rings, which proves their crystalline nature. X-ray diffraction analyses have shown that their concentration is generally higher in the 5- to 10-Am-thick outer layer than in the cortex [32]. These observations could be interpreted on the basis of hair lipids in the form of calcium soaps, i.e. where the calcium is in a form readily removable by hydrochloric acid. It also follows that without such calcium sequestration the lipid component will be incapable of extraction with an organic solvent. Thus, despite the large number of analyses carried out over the years of lipid extracted from hair by solvents, the soaps may have escaped detection because the calcium had not been first removed. The existence of calcium soaps at the surface of hair was in fact inferred a long time ago from chemical investigations [22]; our observations strongly support the existence of internal calcium soaps in hair. We have not been able to secure information as to the binding in the hair of second type of calcium, i.e. that which
Fig. 3. Microfluorescence images of a section of human scalp hair (high calcium content hair, same subject as for Fig. 2) treated in a 0.1 N hydrochloric acid solution, collected at the Ca Ka and S Ka energies. (a) and (b) give respectively the calcium and sulfur distributions across the section obtained with a 1 1Am2 pixel size (beam spot size, 0.3 0.3 m2). Data collection was 1 s/pixel. After lipid removal by HCl, the remaining calcium atoms are mainly located in and around the wall of the medulla. Note the tear in the sample.
58
C. Me´rigoux et al. / Biochimica et Biophysica Acta 1619 (2003) 53–58
resists removal by hydrochloric acid. However, it is reasonable to assume that they are associated to proteins, which would make them less removable by hydrochloric acid. Calcium ions are in fact involved in many biochemical processes. The existence of such proteins has already been established for hair follicle but not yet for hair shaft, except for an S100 family protein in the cuticle. It is worth mentioning that a link between Ca2 + binding proteins (annexins) and lipids in the sarcolemma of smooth muscle cells was reported: the Ca2 + ions regulate the association of lipid microdomains in the plasma membrane [33]. A similar link could be envisaged for hair, at least in the hair follicle. The origin of the calcium-binding proteins is certainly endogenous. On the contrary, the origin of the calcium atoms bound to lipids can be either endogenous or exogenous [34], even if they are located inside the hair fiber. The high variability of the internal calcium soap concentration from one subject to another does not bring any light on this question. The origin of such huge variability remains unclear; further studies are needed to investigate this point which could reflect relevant physiological, biological, hair care, pathological or environmental factors.
Acknowledgements We thank Dr. Pierre Chevallier for support in the fluorescence data analysis and the reviewer who suggested the association of calcium-rich granules with nuclear remnants and made valuable suggestions for the discussion about calcium soap extraction.
References [1] [2] [3] [4] [5]
S.B. M’Baku, R.M. Parr, J. Radioanal. Chem. 69 (1982) 171. T.F. Dutcher, S.J. Rothman, J. Invest. Dermatol. 17 (1951) 65. F.G.E. Pautard, Nature 199 (1963) 531. A. MacPherson, J. Basco, Sci. Total Environ. 1 – 3 (2000) 11 – 19. V. Valkovic, I. Fundamentals and Methods for Measurement of Elemental Composition, CRC Press, Boca Raton, FL, 1988, p. 1. [6] B. Stocklassa, M. Aransay-Vitores, G. Nilsson, C. Karlsson, D. Wiegleb, B. Forslind, J. Cosmet. Sci. 52 (2001) 297 – 311. [7] A.J.J. Bos, C.C.A.H.V.D. Stap, W.J.M. Lenglet, R.D. Vis, V. Valkovic, IEEE Trans. Nucl. Sci. 30 (1983) 1249. [8] A.J.J. Bos, R.D. Vis, H. Verheul, M. Prins, S.T. Davies, D.K. Bowen, J. Makjanic, V. Valkovic, Nucl. Instrum. Methods Phys. Res., B 3 (1984) 232.
[9] A.J.J. Bos, R.D. Vis, W.J.M. Lenglet, H. Verheul, V. Valkovic, J. Makjanic, in: P Bra¨tter, P. Shramel (Eds.), Proceedings 2nd Workshop in Trace Element Analytical Chemistry in Medicine and Biology, Walter de Gruyter, Berlin, 1983, p. 787. [10] K.G.K.R. Malmquist, B. Forslind, Nucl. Instrum. Methods 181 (1981) 465. [11] M. Uo, F. Watari, A. Yokoyama, H. Matsuno, T. Kawasaki, Biomaterials 20 (1999) 747 – 755. [12] J.A. Cookson, F.D. Pilling, Phys. Med. Biol. 20 (1975) 1015 – 1020. [13] I. Bogdanovic, M. Jaksic, S. Fazinic, J.W. Hopewell, J.H. Wilkinson, V. Valkovic, Nucl. Instrum. Methods Phys. Res., B 104 (1995) 339 – 343. [14] U. Wollina, R. Klinger, R. Wetzker, R. Reissmann, B. Knopf, Arch. Dermatol. Res. 280 (1989) 497 – 498. [15] F. Matsumura, S. Yamashiro, Curr. Opin. Cell Biol. 5 (1993) 70 – 76. [16] L. Wood, D. Carter, M. Mills, N. Hatzenbulher, G. Vogel, J. Invest. Dermatol. 96 (1991) 383 – 387. [17] R.A. Fava, L.B. Nanney, D. Wilson, L.E. King Jr., J. Invest. Dermatol. 101 (1993) 732 – 737. [18] K. Kizawa, H. Uchiwa, U. Murakami, Biochim. Biophys. Acta 1312 (1996) 94 – 98. [19] A.W. Weitkamp, A.M. Smiljanic, S. Rothman, J. Am. Oil Chem. Soc. 69 (1947) 1936. [20] N. Nicolaides, S. Rothman, J. Invest. Dermatol. 21 (1953) 9 – 12. [21] O. Sakamoto, Y. Fujinuma, T. Ozawa, J. Am. Oil Chem. Soc. 54 (1977) 143. [22] K.V. Curry, S. Golding, J. Soc. Cosmet. Chem. 22 (1971) 681 – 699. [23] J. Susini, R. Barrett, B. Kaulich, S. Oestreich, M. Salome´, in: W. Meyer, A. Warwick, D.T. Attwood (Eds.), X-Ray Microscopy Proceedings of the Sixth International Conference, vol. 507, American Institute of Physics, New York, 2000, pp. 19 – 26. [24] E. Di Fabrizio, F. Romanato, M. Gentili, S. Cabrini, B. Kaulich, J. Susini, R. Barrett, Nature 401 (1999) 895 – 898. [25] C. David, B. Kaulich, R. Barrett, M. Salome´, J. Susini, Appl. Phys. Lett. 77 (2000) 3851 – 3853. [26] N. Abbott, A. Goodings, J. Text. Inst. 40T (1949) 232. [27] R.D.B. Fraser, T.P. MacRae, G.E. Rogers, KERATINS: Their Composition, Structure and Biosynthesis, Charles Thomas, Springfield, IL, 1972. [28] E. Bramanti, F. Ronca, L. Teodori, M.L. Truica, F. Papineschi, G. Spremolla II, P. Vergamini, E.A. Benedetti, J. Soc. Cosmet. Chem. 43 (1992) 285. [29] J.L. Bantignies, G. Fuchs, G.L. Carr, G.P. Williams, D. Lutz, S. Marull, Int. J. Cosmet. Sci. 20 (1998) 381 – 394. [30] L. Kreplak, C. Merigoux, F. Briki, D. Flot, J. Doucet, Biochim. Biophys. Acta 1547 (2001) 268 – 274. [31] R.D.B. Fraser, T.P. Macrae, G.E. Rogers, B.K. Filshie, J. Mol. Biol. 7 (1963) 90 – 91. [32] B. Busson, P. Engstro¨m, J. Doucet, J. Synchrotron Radiat. 6 (1999) 1021 – 1030. [33] E.B. Babiychuck, A. Draeger, J. Cell Biol. 150 (2000) 1113 – 1123. [34] J. Basco, L. Sarkadi, E. Koltay, Int. J. Appl. Radiat. Isot. 33 (1982) 5.