This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Lyubimova, N V Articles by Martin, R F Search for Related Content PubMed PubMed Citation Articles by Lyubimova, N V Articles by Martin, R F

In vivo radioprotection of mouse brain endothelial cells by Hoechst 33342

¹ Trescowthick Research Laboratories
² Statistical Centre, Peter MacCallum Cancer Institute, St Andrews Place, Melbourne, Victoria 3000, Australia

	Abstract

Top Abstract Introduction Materials and methods Results, statistical analysis... References

 
Radiation-induced loss of mouse brain endothelial cells has been examined in mice given an intravenous injection of the DNA-binding radioprotector Hoechst 33342 (80 mg kg^-1). At the time of irradiation, 10 min after injection, Hoechst fluorescence in the brain was confined to the endothelial cells. Endothelial cell density was measured using a histochemical fluorescence technique that had been used previously to monitor post-irradiation changes in endothelial cell density in rat brain, in which it was shown that a sensitive subpopulation comprising about 15% of the endothelial cells was lost within 24 h of radiation exposure. The present study shows a similar dose–response for the control mice, with depletion of the sensitive subpopulation to 85% being almost complete after a dose of 2.5 Gy -rays. However, in mice irradiated 10 min after Hoechst 33342 administration, doses between 12 Gy and 20 Gy were required to ablate these cells. The kinetics of cell loss and the rather large dose modification factor suggests that Hoechst 33342 may be suppressing an apoptotic response in this subpopulation. Whatever the mechanism involved, Hoechst 33342 clearly provides substantial protection against early radiation-induced endothelial cell loss. Further studies are necessary to determine the extent to which this initial protection translates into an improved long-term survival of the "protected" cells and, especially, to see whether this endothelial cell protection can ameliorate the later consequences of central nervous system irradiation, namely necrosis and paralysis.

	Introduction

Top Abstract Introduction Materials and methods Results, statistical analysis... References

 
The minor-groove-binding DNA ligand Hoechst 33342 and its analogues are known to protect DNA and cells from ionizing radiation [1]. Radioprotective activity was first described by Smith and Anderson [2] who reported a dose modification factor (DMF) of 1.5 at the 10% survival level in HT29 human adenocarcinoma cells exposed to 8.7 µM Hoechst 33342. Subsequent studies have reported higher [3] and lower [4] DMFs, this variation probably reflecting differences between cell lines and experimental conditions. However, it is clear that Hoechst 33342 is more potent than classical aminothiol radioprotectors: similar DMFs can be achieved with 100-fold lower concentrations of Hoechst 33342 than, for example, WR 1065 [5]. This high potency of Hoechst 33342 could reflect an advantage of directly targeting radioprotectors to DNA [6]. Mechanistic studies have shown that bound ligand can protect isolated DNA from radiation-induced strand breaks [4] and it is presumably this decrease in initial DNA damage that accounts for radioprotective activity at the cellular level. Although in an early report Young and Hill [3] failed to detect radioprotective activity in vivo, radioprotection has now been demonstrated in mice, both for whole body exposure after intravenous Hoechst 33342 at doses as low as 2 mg kg^-1 [7] and in lung after 80 mg kg^-1 [8]. In the latter case, the DMF was 1.2 and close to that reported for WR2721 in a similar study of mouse lung [9].

Limited penetration through cell layers is a feature of DNA ligands, which together with the fluorescent properties of Hoechst 33342 has been exploited in the use of the ligand as a perfusion marker, both in vitro [10] and in vivo [11, 12]. A good example is provided by a study of rat aorta [13] where, after intravenous injection of Hoechst 33342, uptake was first evident in endothelial cells and then, at increasing times after injection, there was progressive penetration to deeper cell layers. This feature of DNA ligands suggests that by changing the timing of irradiation relative to Hoechst 33342 administration, different cell populations within a tissue could be preferentially protected. For example, this approach could help resolve the question (see [14–16]) of the relative importance of glial and endothelial cell damage in the development of late radionecrosis in the central nervous system (CNS), since systemic administration of a DNA-binding radioprotector such as Hoechst 33342 shortly before irradiation should protect preferentially the endothelium.

If radiation damage to the endothelial cell population underlies the late expression of CNS damage, then radiation-induced changes in endothelial cell numbers might provide a useful and early indication of the likely severity of late reactions. An assay developed by one of us for monitoring radiation-induced changes in endothelial cellularity in rat brain has provided some support for this view [17]. The method was based on the in situ accumulation of dopamine in brain endothelial cells that takes place after treatment of animals with L-DOPA and a monoamine oxidase inhibitor. Dopamine can be converted to a fluorophore by paraformaldehyde exposure, enabling endothelial cell numbers to be quantitated by fluorescence microscopy. In applying this technique in a study of blood vessel length and endothelial cellularity in irradiated rat brain, it was found that although there was no significant change in blood vessel length, there were complex changes in endothelial cell numbers [17]. Thus, after a single dose of 25 Gy 200 kV X-rays, within 1 day endothelial cell numbers were reduced to about 85% of control values. There was a subsequent continuing slow cell loss until endothelial cellularity reached a nadir at about 60% of control density by 6 months post irradiation. A transient repopulation was then seen before a second wave of cell loss at around 15 months after exposure. This second collapse in endothelial cellularity took place at a similar time to the development of brain necrosis and is consistent with the hypothesis that endothelial damage plays a part in late tissue damage.

The initial rapid fall to about 85% of control numbers within 1 day was examined further and found to be dose independent over an exposure range of 5–100 Gy X-rays [17]. This suggested that there is an endothelial subpopulation comprising cells liable to rapid loss (presumably through apoptosis) after relatively low radiation doses. Since the response of these cells might provide a useful early marker to assess the effects of radioprotective agents, it was decided to undertake the present study of the effects of Hoechst 33342 on early post-irradiation endothelial cell loss before investigating the effects on late CNS end-points such as necrosis or paralysis.

	Materials and methods

Top Abstract Introduction Materials and methods Results, statistical analysis... References

 
Male BALB/C mice were used at 8–9 weeks of age and were assigned to control or protected groups. At 10 min before irradiation, the mice in the protected group were injected intravenously with Hoechst 33342 (80 mg kg^-1 at 4 mg ml^-1 in 10% dimethylsulphoxide (DMSO) in saline). For the main series of experiments, unanaesthetized mice were irradiated with single whole body doses between 1 Gy and 25 Gy at a dose rate of 80 cGy min^-1 using Gammacell ¹³⁷Cs and the mice sacrificed 24 h later by decapitation under fluorothane anaesthesia. A smaller experiment was also carried out to confirm that the DMSO present in the carrier had no significant radioprotective effect.

The same technique developed for the earlier study [17] was used to measure endothelial cell density. Two intraperitoneal injections were given to induce a build up of catecholamines (dopamine) in brain endothelial cells. First, iproniazide (1-isonicotinoyl-2-isopropylhydrazide at 100 mg kg^-1) was given 6 h before sacrifice to inhibit endothelial monoamineoxidase activity and, subsequently, the dopamine precursor L-DOPA (3,4-dihydro-1-ß-phenylalanine at 100 mg kg^-1) was given 15 min before decapitation. The brain was then rapidly removed and pieces of tissue between the olfactory bulb and optic chiasma were dissected and immediately frozen in liquid nitrogen. The interval between decapitation and immersion of the tissue samples into liquid nitrogen was 1–2 min. Frozen samples were maintained at sub-zero temperatures during lyophilization for 3 days and then exposed to formaldehyde vapour at 70% humidity and 80 °C to convert the accumulated catecholamines to fluorescent isoquinolines. Brain pieces were directly embedded in paraffin under vacuum, sectioned at 10 µm and mounted on dry slides in immersion oil. Scoring was carried out by fluorescence microscopy using a bandpass filter that obviated the need to use a monochromatic laser source by detecting isoquinoline fluorescence and excluding fluorescence of Hoechst 33342. Three sections were assessed for each animal; in each section of cortex (grey matter), endothelial cell numbers were counted over an area covering all cortical layers and comprising 30 graticule fields, each field corresponding to 0.0625 mm². Data from at least three mice were used for each treatment group. The mean cell count for each mouse was calculated for each treatment group and an overall mean for each dose/group was obtained from the mean counts for each mouse in the group.

All the regression analyses described in the results section were performed using the SPSS statistical package (SPSS Advanced Statistics 7.5, SPSS Inc., Chicago, IL). Analysis of the results of the DMSO experiments used the StatXact package, StatXact 3 for Windows (CYTEL Software Corporation, MA).

	Results, statistical analysis and discussion

Top Abstract Introduction Materials and methods Results, statistical analysis... References

 
The dose-related changes to mean cell counts (±standard errors) are plotted for each treatment group in Figure 1. For animals given radiation alone, there was a loss of up to 15% of the endothelial cell population, similar to the pattern previously reported for rats [17]. However, compared with the earlier study, a wider range of low doses was given and, as can be seen from the figure, loss of the sensitive subpopulation was already substantial after a dose of 2.5 Gy. In animals pre-treated with Hoechst 33342, loss of endothelial cells was only evident at radiation doses greater than 10 Gy.

View larger version (15K): [in this window] [in a new window]   Figure 1. Dose–response for rapid radiation-induced reduction of endothelial cell numbers and the effect of prior administration of Hoechst 33342. Endothelial cell numbers were scored 24 h after irradiation, as described in the text. Some mice were injected intravenously with Hoechst 33342 10 min prior to irradiation () while the controls (•) were exposed to radiation only. Data points are mean cell counts for each group of mice. Standard errors are also shown (bars). The lines (—, protected fit; --, control fit) indicate the results of non-linear regression analysis using Equation (1) (see text and Table 1 for details).

where Y is the cell count, D is the dose and E, Er, and ß are parameters to be estimated. E represents the cell count in unirradiated control animals and Er is the asymptotic minimum of the cell count as the dose tends to infinity and only cells in a non-affected subpopulation remain. The dose that leads to a 50% reduction in the radiosensitive population of cells (ED₅₀) is represented by in the model. The parameter ß reflects the steepness of the dose–effect curve. The parameters were estimated from a non-linear regression on the overall means after weighting by the numbers of mice. All regressions and parameter values were derived with E and Er constrained to common values for the control and protected groups.

To determine the significance of the differences in cell counts between the protected and control groups, a regression was first fitted to all the data (both from control and protected groups). Second, the data were fitted so that independent values of and ß were estimated for the control and protected data. The following test statistic was then used:

where R₁², R₂², ₁ and ₂ are the residual sum of squares and degrees of freedom associated with first and second fits, respectively. This statistic was compared with the F distribution with ₁-₂ and ₂ degrees of freedom. A highly significant result was obtained (p<0.00001) for the difference between the control and protected groups. The independent fits for and ß are given in Table 1. The estimated ED₅₀±standard error for the protected and the control groups were 18.5±4.0 and 6.3±6.1, respectively. The fitted curve from the regression for each group is also displayed in Figure 1. Clearly, treatment with Hoechst 33342 provided a very substantial protective effect, the data indicating a DMF of around 3.

View this table: [in this window] [in a new window]   Table 1. Parameters derived from the curve fitting displayed in Figure 1

The use of DMSO in saline as the carrier for the Hoechst 33342 should be mentioned, since DMSO itself possesses radioprotective properties. For the present study, DMSO was not an essential part of the experimental design but was included since this study is part of a programme involving comparison of radioprotective properties in a range of newly synthesized bisbenzimides. Some of these compounds are less water soluble than Hoechst 33342 and inclusion of 10% DMSO in the delivery vehicle is helpful. However, we had indirect evidence that intravenous delivery of this volume of DMSO in saline was unlikely to have a significant radioprotective effect from experiments investigating radioprotection of mouse lung by a new bisbenzimide, as in this tissue the putative cellular targets include endothelial cells. The radiation dose–response for loss of lung function for mice pre-treated with intravenous vehicle alone (10% DMSO in saline) was indistinguishable from that for intact controls (paper in preparation). Moreover, the question of the protective effects of DMSO was investigated directly in a separate experiment that compared the effect of a 9 Gy irradiation on cellularity at 24 h in four intact mice and in three mice injected with DMSO/saline vehicle. To compare cell counts in these two groups, a permutation test was carried out owing to the small number of mice available for analysis. There was no statistically significant difference in the cell counts for the two groups (p=0.46). The mean (±standard error) endothelial cell densities for the irradiation only controls and for the irradiated–DMSO groups were 342.3±2.4 cells mm^-2 and 345.9±4.5 cells mm^-2, respectively.

Classical radioprotectors, which act by scavenging hydroxyl radicals, generally confer radioprotection characterized by DMF values of 2 to 3. This limit reflects the relative contributions of indirect (radical-mediated) and direct action. Recent pulse radiolysis studies support the hypothesis that Hoechst 33342 acts by chemical repair of radiation-induced radical species on DNA [18] and, since some of those species are common to both direct and indirect action [19], this could provide a basis for the relatively high radioprotection observed here. However, the DMFs reported for Hoechst 33342 in experiments with cultured cells are not particularly high; they are generally less than 2.

In seeking other possible explanations for the large DMF found with Hoechst 33342 in the present experiments, the mode of cell death of the radiosensitive subpopulation needs to be considered. Endothelial cells have very low proliferation rates, corresponding to population replacement times of the order of months [20], so the expected normal daily replacement rate is less than 1% [21]. Thus, the loss of a subpopulation of some 15% of these cells within a day must involve interphase death, presumably by apoptosis, although we have yet to confirm this directly. There are reports of apoptosis in cerebral endothelial cells [22]. Since DNA damage recognition is likely to be an early event in radiation-induced apoptosis, repair of initial DNA damage by Hoechst 33342 would be expected to also confer protection against both apoptosis and interphase death. However, it is possible that the DNA lesions recognized in the apoptotic response are chemically different from those responsible for reproductive cell death, with different susceptibility to chemical repair.

There are a number of other factors that could increase the DMF beyond the change due to direct repair by the ligand. The DMF would also be affected were Hoechst 33342 to interfere at other points downstream in the signal transduction pathway(s) for apoptosis. This latter scenario raises the possibilities that the availability of the ligand shortly after irradiation could provide some protection, or that the time course of the apoptotic response could be delayed. Such a delay would increase the opportunity for repair since there is competition between apoptotic and repair pathways. Indeed, delaying the apoptotic response has been shown to produce real increases in cell survival after irradiation [23]. While it might be possible that the vasoconstrictive effects of Hoechst 33342 at high dose could cause transient hypoxia, this seems unlikely as radioprotection has been reported at Hoechst 33342 doses too low to cause vasoconstriction [7]. Moreover, in a privileged site such as the brain, serious hypoxia is improbable.

Whatever the explanation for the high DMF observed, it must be emphasized that it has been shown to apply only to the smaller subpopulation of endothelial cells comprised of cells characterized only by their sensitivity to an early apoptotic death and with an as yet undefined role. The high DMF may well not fully extend to all endothelial cells or to late necrosis. Nevertheless, it cannot be assumed that endothelial cells in the unaffected subpopulation are free of radiation injury and, in so far as protection by Hoechst 33342 results from its effects on DNA damage, at least some protection should extend to these endothelial cells, although quantification as to how much is beyond the scope of the present investigation. An investigation of the effectiveness of Hoechst 33342 in protecting against late necrosis is clearly warranted.

It is interesting to compare the results of this and an earlier study [17], both of which infer a role for radiation-induced apoptosis in late CNS necrosis, with a recent report by Li et al [24] involving direct observation of apoptosis in adult rat spinal cord. While the present study focused on endothelial cells, Li et al [24] found apoptosis in glial cells but not in vascular endothelial cells. At 8 h post irradiation, a little less than 15% of glial cells were apoptotic. This result is clearly significant, but with the low incidence of apoptosis the importance of the timing of observations is critical and there is a possibility that apoptosis in the endothelial cells could have been missed. As noted previously, recent in vitro studies of endothelial cells of various origins have reported apoptosis [22]. The present experimental approach, which measures loss of cells, differs from that of Li et al [24] in being a cumulative endpoint rather than a snapshot. More generally, two conclusions can be drawn from the comparison of our results with those of Li et al [24]. One is that it would not be prudent to cite the results of Li et al [24] as evidence that endothelial cells are not targets for radiation effects on the CNS. Second, the comparison raises the question of possible radiobiological differences between brain and spinal cord, which could be addressed by seeking direct evidence of apoptosis in brain.

When originally considering the radioprotection of endothelial cells by Hoechst 33342, a potential problem was thought to be the P-glycoprotein pump activity known to be prevalent in these cells [25]. Indeed, in primitive haemopoietic cells exposed to Hoechst 33342, P-glycoprotein pump activity provides a means of enriching stem cell populations by fluorescence-activated cell sorting, as the stem cells are characterized by very low fluorescence [26]. However, our results clearly demonstrate that P-glycoprotein pump activity does not prevent sufficient Hoechst 33342 being taken up into endothelial cells to modify radiation responses (Figure 1). This effective uptake into the endothelial cell population, in combination with the possibility of exploiting the perfusion kinetics of DNA ligands such as Hoechst 33342, could facilitate investigations into the role of endothelial cells in normal tissue responses to radiation.

In summary, this study confirms, first, that Hoechst 33342 can act as a radioprotector in vivo. Second, protection of a radiosensitive subpopulation of brain endothelial cells, probably killed by apoptosis, suggests that Hoechst 33342 can substantially modify the apoptotic response. In so far as there could be a role for radiation-induced apoptosis of endothelial cells in the development of late CNS damage, we conclude that follow-up investigations of the effects of Hoechst 33342 on apoptosis and also on late radionecrosis and paralysis would be particularly worthwhile. Finally, this study indicates that the pharmacokinetic properties of DNA-binding radioprotectors may prove helpful in elucidating the roles of endothelial and parenchymal damage in normal tissue responses to ionizing radiation.

	Footnotes

 
This project was supported by the Australia/Russia Cooperation in Medical Science and Public Health

Current address for Dr N V Lyubimova: Institute of Pathology, University of Bern, Murtenstrasse 31, Bern 3010, Switzerland

Received for publication January 4, 1999. Revision received July 14, 2000. Accepted for publication September 25, 2000.

	References

Top Abstract Introduction Materials and methods Results, statistical analysis... References

 

1. Martin RF, DNA-binding bibenzimidazoles as radioprotectors. In: Bump EA, Malaker E, editors. Radioprotectors: chemical, biological and clinical perspective. Boca Raton, FL: CRC Press Inc, 1998:151–66.
2. Smith PJ, Anderson CO. Modification of the radiation sensitivity of human tumour cells by a bis-benzimidazole derivative. Int J Radiat Biol 1984;46:331–44.
3. Young SD, Hill RP. Radiation sensitivity of tumour cells strand in vitro or in vivo with the benzimidazole fluorochrome Hoechst 33342. Br J Cancer 1989;60:715–21.[Medline]
4. Denison L, Haigh A, D'Cunha G, Martin RF. DNA ligands as radioprotectors: molecular studies with Hoechst 33342 and Hoechst 33258. Int J Radiat Biol 1992;61:69–81.[Medline]
5. Murray D, Prager A, Milas L. Radioprotection of cultured mammalian cells by the aminothiols WR-1065 and WR-255591: correlation between protection against DNA double-strand breaks and cell killing after -radiation. Radiat Res 1989;120:154–63.[Medline]
6. Zheng S, Newton GL, Gonick G, Fahey RC, Ward JF. Radioprotection of DNA by thiols: relationship between the net charge on a thiol and its ability to protect DNA. Radiat Res 1988;114:11–27.[Medline]
7. Singh SP, Jayanth VR, Chandna S, Dwarakanath BS, Singh S, Adhikari JS, et al. Radioprotective effects of DNA ligands Hoechst-33342 and 33258 in whole body irradiated mice. Indian J Exp Biol 1998;36:375–84.[Medline]
8. Martin RF, Broadhurst S, DeAbrew S, Budd R, Sephton R, Reum M, et al. Radioprotection by DNA ligands. Br J Cancer 1996;74:S99–S101.
9. Travis EL, De Luca BS. Protection of mouse lung by WR-2721 after fractionated doses of irradiation. Int J Radiat Oncol Biol Phys 1985;11:521–6.[Medline]
10. Durand RE. Use of Hoechst 33342 for cell selection from multicell systems. J Histochem Cytochem 1982;30:117–22.[Abstract]
11. Olive PL, Chaplin DJ, Durand RE. Pharmacokinetics binding and distribution of Hoechst 33342 in spheroids and murine tumours. Br J Cancer 1985;52:739–46.[Medline]
12. Chaplin DJ, Olive PL, Durand RE. Intermittent blood flow in a murine tumour: radiobiological effects. Cancer Res 1987;47:597–601.[Abstract/Free Full Text]
13. Daly CJ, Gordon JF, McGrath JC. The use of fluorescent nuclear dyes for the study of blood vessel structure and function: novel applications of existing techniques. J Vasc Res 1992;29:41–8.[Medline]
14. Myers R, Rogers M, Hornsey S. A reappraisal of the roles of glial and vascular elements in the development of white matter necrosis in irradiated rat spinal cord. Br J Cancer 1986;53(Suppl. 7):221–3.
15. Hopewell JW, Campling D, Calvo W, Reinhold HS, Wilkinson JH, Yeung TK. Vascular irradiation damage: its cellular basis and likely consequences. Br J Cancer 1986;53(Suppl. 7):181–91.[Medline]
16. van der Kogel AJ. Radiation-induced damage in the central nervous system: an interpretation of target cell responses. Br J Cancer 1986;53(Suppl. 7):207–17.
17. Lyubimova NV, Levitman MKh, Plotnikova ED, Eidus LKh. Endothelial cell population dynamics in rat brain after local irradiation. Br J Radiol 1991;64:934–40.[Abstract/Free Full Text]
18. Martin RF, Anderson RF. Pulse radiolysis studies indicate that electron transfer is involved in radio-protection by Hoechst 33342 and methylproamine. Int J Radiat Oncol Biol Phys 1998;42:827–31.[Medline]
19. O'Neill P, Chapman PW. Potential repair of free radical adducts of dGMP and dG by a series of reductants. A pulse radiolysis study. Int J Radiat Biol 1985;47:71–80.
20. Hobson B, Denekamp J. Endothelial proliferation in tumours and normal tissues: continuous labeling studies. Br J Cancer 1984;49:405–13.[Medline]
21. Korr H, Schultz B, Maurer W. Autoradiographic investigations of glial cell proliferation in the brain of adult mice. II. Cycle time and mode of proliferation in neuroglia and endothelial cells. J Comp Neurol 1975;160:477–90.[Medline]
22. Langley RE, Bump EA, Quartuccio SG, Medeiros D, Braunhut SJ. Radiation-induced apoptosis in microvascular endothelial cells. Br J Cancer 1997;75:666–72.[Medline]
23. Radford IR. Phorbol esters can protect mouse pre-T cell lines from radiation-induced rapid interphase apoptosis. Int J Radiat Biol 1994;65:345–55.[Medline]
24. Li YQ, Guo YP, Jay V, Stewart PA, Wong CS. Time course of radiation-induced apoptosis in the adult rat spinal cord. Radiother Oncol 1996;39:35–42.[Medline]
25. Beaulieu E, Demeule M, Pouliot JF, Averill-Bates DA, Murphy GF, Beliveau R. P-glycoprotein of blood brain barrier: cross-reactivity of Mab C219 with a 190 kDa protein in bovine and rat isolated brain capillaries. Biochim Biophys Acta 1995;1233:27–32.[Medline]
26. Wolf NS, Kone A, Priestley GV, Bartelmez SH. In vivo and in vitro characterisation of long term repopulating primitive haematopoietic cells isolated by sequential Hoechst 33342-rhodamine 123 FACS selection. Exp Hematol 1993;21:614–22.[Medline]

This article has been cited by other articles:

				N Lyubimova and J W Hopewell Experimental evidence to support the hypothesis that damage to vascular endothelium plays the primary role in the development of late radiation-induced CNS injury Br. J. Radiol., June 1, 2004; 77(918): 488 - 492. [Abstract] [Full Text] [PDF]

				A. Adhikary, V. Buschmann, C. Muller, and M. Sauer Ensemble and single-molecule fluorescence spectroscopic study of the binding modes of the bis-benzimidazole derivative Hoechst 33258 with DNA Nucleic Acids Res., April 15, 2003; 31(8): 2178 - 2186. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Lyubimova, N V Articles by Martin, R F Search for Related Content PubMed PubMed Citation Articles by Lyubimova, N V Articles by Martin, R F