Journal of Microscopy, Vol.166, I't 2, M a y 1992, p p . 199-205. Received 13 September 1991; revised and accepted 11 November 1991
Staining sections of water-miscible resins
2. Effects of staining-reagent lipophilicity on the staining of glycol-methacrylate-embedded tissues by R I C H A R D w. HOROBIN*, P E T E R 0. G E R R I T S t and D A V I D J. WRIGHT*', *Department of Biomedical Science, The University, Shefield S10 2 T N , U.K. and tAnatomy and Embryology Laboratory, The University of Groningen, Oostersingel69, 97 13 E Z Groningen, The Netherlands
w O R D S . Artefact, glycol methacrylate (GMA), lipophilicity, structure-staining correlations, structure-activity relationship, water-miscible resin.
KEY
SUMMARY
Glycol methacrylate (GMA) sections of animal tissues were stained with a group of twenty-seven reagents of very varied chemical characteristics. T h e artefactual background staining of the resin was found to be dependent on the hydrophilic/lipophilic character of the staining reagent, as estimated from the logarithm of its octanol-water partition coefficient (log P). Intense background staining occurred with lipophilic stains, whose log P > 2 . I n keeping with this, use of G M A semi-permeable membranes for enzyme histochemistry failed to give staining when using a lipophilic substrate, probably because the substrate was trapped in the membrane. An analysis of other routine histochemical stains-in terms of the probable occurrence of high resin background staining and low tissue sensitivity-is made. A numerical guide is provided to help avoid artefacts resulting from hydrophobic and size effects. Note: small, hydrophilic reagents (log P < 0; molecular weight < 550 Da) are least likely to show either type of artefact. Conversely, reagents which are lipophilic, or/and of intermediate size (log P z 2; 550 < ionic weight < 1000 Da), give strong background staining. INTRODUCTION
T h e use of resin-embedded sections, especially of glycol methacrylate (GMA), is now commonplace in light microscopy. T h e advantages of using GMA include ease of cutting thin sections, even of hard tissues, and the applicability of staining procedures developed for paraffin and cryostat sections. However, such staining methods may fail with GMA sections, and may give high background staining of the resin itself. We have sought to clarify such problems, and were initially concerned with the influences of reagent molecular size (Gerrits et al., 1990). I n this earlier study all the reagents used were hydrophilic, since a pilot investigation indicated that lipophilic dyes gave high resin background staining. It is such background lipophilia which are the subject of this present paper.
'
Present address. Department of Zoology, 1 1 17 West Johnson, The University of Wisconsin, Madison, WI 53706, USA. (
1992 The Royal Microscopical Society
199
200
R. W . Horobin, P. 0. G e r r i t s and D . J . Wright
Since we wished to evaluate the role of stain-resin hydrophobic bonding, it was necessary to exclude other affinity or rate effects. Artefactual staining of the resin by basic dyes was avoided by using methacrylic acid-free resin (see Gerrits & van Leeuwen, 1987). In choosing staining reagents, effects due to size-dependent rate of penetration had to be avoided. Consequently, most reagents used in the present study are small, with ionic weights < 550 Da. However, reagents of moderate size can give useful information when they fail to stain the resin, although with such reagents resin staining is ambiguous, since it could be due to the effects of the rate of stain penetration (Gerrits et al., 1990). With these issues in mind, stain-resin hydrophobic bonding was investigated using staining reagents which varied from very hydrophilic to markedly lipophilic. It was supposed that background staining of the resin due to hydrophobic bonding would be most intense with lipophilic compounds. T h e hydrophilicity-lipophilicity of the reagents was modelled numerically using the log P parameter, i.e. the logarithm of the octanol-water partition coefficient (Hansch & Leo, 1979). MATERIALS AND METHODS
Preparation of resin sections for staining with dyes Various rat tissues were fixed for 24 h in 4O, neutral buffered aqueous formaldehyde. Tissues were dehydrated through the ethanols, and embedded in G M A using the JB4 kit (Polyscience, Warrington, U.S.A.), according to their Datasheet No. 123. Sections were cut at 2 pm with a glass knife, on an Anglia Scientific AS 500 microtome. Sections were floated onto water at room temperature, mounted on slides, and dried down at 37°C for 2 h or at 60-C for 10 min. Preparation of resin sections for enzyme histochemistry This was carried out in a similar way, as described in detail in Gerrits et al. (1989). This involved covering the resin tissue with a 2-pm section of tissue-free GMA, as a semi-permeable membrane. T h e dyes used are listed in Table 1, together with their sources and certain physicochemical characteristics. T h e log P values were obtained by calculation; the procedure is described in detail in Hansch & Leo (1979). T h e purity of the dyes was assessed using the reverse-phase thin-layer chromatography system of Proctor & Horobin (1985). Staining of resin sections Staining solutions were made up by dissolving 0.1 g of dye in 100 ml of aqueous buffer. T h e buffers used were 0.1 M p H 4 acetate, 0.1 M p H 7 phosphate, and 0.1 M p H 9 barbitone. Non-ionic dyes were used at p H 7; most acid and basic dyes were used at both p H 4 and 9. T h e only exception to this was Aldehyde Fuchsin, which was prepared following the procedure of Wolters et al. (1983), and was thus dissolved in acid alcohol usual for this reagent. Sections were stained until no further increases in coloration were seen, washed in running water for 1 min, and blotted or blow dried with compressed air. Sections were then air dried for 20 min, dipped in xylene, and mounted in a synthetic resin (DPX). E n z y m e histochemistry on G M A sections Non-specific esterase (i.e. 1-naphthyl acetate esterase) was demonstrated using diazotized Pararosanilin as the coupler (Bancroft & Stevens, 1982). Alkaline phosphatase was demonstrated using Naphthol AS-MX phosphate as the substrate and Fast Blue B as the visualizing agent (Gerrits et aZ., 1989). Acid phosphatase was demonstrated using 1-naphthyl phosphate as the substrate, and diazotized Pararosanilin as the coupler (after Barka & Anderson, 1963).
Lipophilic reagents and water-miscible resins
20 1
Table 1. Reagents used for staining GMA sections.
Reagent (Colour Index identification number) Acetoquinone Blue BF (61545) Astrazone Orange R (48040) Azo Fuchsin B (16550) Biebrich Scarlet (26905) Bismark Brown R (21010) Brazilein (-) Chromotrope 2R (16570) Crystal Violet (42555) Cuprolinic Blue (-) Dispersol Blue 7G (62500) Eosin Y (45380) Ethyl Violet (42600) Fast Blue BB (37175) Hematein (-) Hexazotized Pararosanilin (-) Indanthrene Bordeaux B (60705) Janus Green B (1 1050) Mordant Brown 6 ( 11875) Naphthol AS-BI phosphate (-) Naphthol AS-MX phosphate (-) 1-Naphthyl acetate (-) 1-Naphthyl phosphate (-) Orange G (16230) Purpurin (75410) Rhodamine 6G (45 160) Thionin (52000) Toluidine Blue (52040)
log P*
+ 0.2 + 2.2 - 5.0
- 3.2 - 0.4
+ 1.3 - 4.3 + 1.2 - 14.5 + 1.3 + 1.4 + 3.4 - 0.4 + 1.2 < -5 + 0.9 + 0.4 - 0.4 + 0.3 i 1.4 + 3.4 - 1.0 - 5.0
+
+ 1.0 3.5
- 2.1 - 0.6
Ionic or molecular weight (Da) 326 39 1 436 510 390 286 422 372 640 358 646 456 312 300 312 498 475 316 294 369 167 222 406 256 443 225 270
Charge 0 +1 -2 -2 +l 0
-2 +1 +4 0 -2 +1 +1 0 +3 0 +1 -1 -1 -1 0 -1 -2 0 +1 +1 +1
Stain sourcet EG EG EG EG AL EG EG EG BD MF AL KL SI KL AL EG BD AL SI SI SI SI AL EG AL AL AL
The log of the octanol-water partition coefficient. BD = British Drug Houses, E G = Edward Gurr, K L = Koch Light, M F = manufacturer's sample.
t AL = Aldrich,
Assessment of relative staining intensities This was carried out subjectively by three independent observers. RESULTS
Overall summary of background resin staining Background staining of resin in sections may be summarized as follows. (1) T h e following dyes gave no staining of the resin: Azo Fuchsin B (sample l), Chromotrope 2R, Cuprolinic Blue (acidic solution), Orange G, Thionin. (2) T h e following dyes gave a trace of resin staining when viewed macroscopically, but this was not noticed in the microscope: Azo Fuchsin B (sample 2 ) , Biebrich Scarlet, Bismark Brown, Cuprolinic Blue (alkaline solution), Mordant Brown 6, Toluidine Blue. ( 3 ) T h e following dyes stained the resin weakly but sufficiently to be noticed in the microscope: Acetoquinone Blue BF, Brazilein, Dispersol Blue 7G, Eosin Y, Hematein, Indanthrene Bordeaux B, Purpurin. (4) T h e following dyes stained the resin intensely, obscuring the specimen when viewed microscopically: Astrazone Orange R, Crystal Violet, Ethyl Violet, Janus Green B, Rhodamine 6G.
202
R. W . Horobin, P. 0. Gerrits and D . J . Wright
% 0 0
8 w
0 I
none
trace
weak
intense
Degree of resin background staining Fig. 1. A summary of the influence of the hydrophilic lipophilic character of stains on the degree of resin background staining.
Fig. 2. An illustration of the effect of reagent lipophilicity on resin background staining. Section (a) was stained with the hydrophilic basic dye Thionin, whilst section (b) was stained with the lipophilic basic dye Janus Green B. Arrows indicate regions of tissue-free resin: black arrows, unstained resin; white arrows, stained resin. Magnification x 550.
Lipophilic reagents and water-miscible resins
203
Enzyme histochemical staining. Acid and alkaline phosphatases were successfully demonstrated using the GMA-membrane technique, whereas non-specific esterase was not. Influences of hydrophobic effects on the background resin staining of sections As expected (Tanford, 1980), hydrophobic effects increased as the lipophilicity of the staining reagents rose, as seen in Fig. 1. Since, for practical purposes, background staining not noticeable in the microscope may be ignored, Fig. 1 suggests the following rules of thumb. Small molecular weight dyes with log P < 0 will usually give minimal or no hydrophobic background staining; dyes with log P > 2 will often give gross background coloration of resin; dyes of intermediate lipophilicity give weak staining of the resin. These conclusions are illustrated by the micrographs seen in Fig. 2. Background staining can of course occur for reasons other than hydrophobic effects, as described in the Introduction. Reagent-resin absorption due to hydrophobic effects will still occur with colourless staining reagents. However, the observed outcomes are now likely to take the form of understaining of the tissue, rather than overstaining of the resin. An example of this was provided by comparing our successful demonstration of acid and alkaline phosphatases with our failure to demonstrate esterase. All the enzyme substrates and visualizing agents used were small, with ionic weights < 500 Da, so staining failure would not have been a size effect (cf. Gerrits et al., 1990). However, log P values of the enzyme substrates differed markedly. T h e ineffective substrate was by far the most lipophilic (log P = 3.4), whilst effective substrates had log P values of - 1.0 and 1.4. It may be supposed that the amount of the more lipophilic substrate reaching the tissue was markedly reduced due to binding to the resin of the overlying membrane. T h e log P cut-off point for effective enzymatic staining is consistent with the results seen with the dyes.
+
+
DISCUSSION
Implications of hydrophobic effects for everyday microscopy Excessive background staining of resin, or understaining of tissue, due to reagent lipophilicity has marked implications for the everyday use of G M A sections. We must ask: what routine staining reagents might give staining artefacts due to the hydrophobic effect? This may be put numerically, since the rule of thumb given above suggests that significant problems may occur with reagents having log P values > 1. Considering dyes and fluorochromes, many anionic compounds are hydrophilic. However, the classic acid dye Eosin Y is slightly lipophilic, and gives sufficient background staining to require an alcohol rinse. T h e commonly used basic dyes and cationic fluorochromes are more often lipophilic. Thus background staining is substantial with Aldehyde Fuchsin and Crystal Violet, and may be expected with reagents such as Granular Blue or Pyronin B. T h e same is true for non-ionic dyes and probes, e.g. Nile Red or the Sudan dyes. However, it is sometimes possible for such reagents to stain lipids in GMA sections, by use of careful differentiation in a plasticizing solvent (cf. Gerrits et al., 1987). T h e most commonly used reactive reagent, Schiff’s, is hydrophilic and gives no problems. T h e same applies to the calcium complexing agent Alizarin Red. However, the reactive thiol reagent Mercury Orange is lipophilic and can give high resin backgrounds, whose removal requires extraction with plasticizing solvents. Routinely used enzyme substrates include several which are lipophilic, with log P values > 1. These include certain naphthyl and indoxyl compounds and their halogenated derivatives, e.g. those used for cholinesterase, esterase and lipase
204
R. W . Horobin, P. 0. Gerrits and D . J . W r i g h t
this zone was not investigated
0
logP
2 r*
Reagent hydrophilicity/lipophilicity Fig. 3. A guide for choosing artefact-free staining reagents for GMA sections. Reagents falling into zone 1 will give few problems. Those in zone 2 will give some resin background, removable with a plasticizing solvent such as ethanol. Reagents falling into zone 3 give no background, but only stain resin-free or surfaceexposed tissue components. Reagents in zone 4 give intense resin background staining.
demonstrations. Peroxidase substrates, e.g. naphthols and arylamines, are also generally lipophilic. Such compounds are used both in enzyme histochemistry and in enzyme-labelled immunostaining.
Problems due to impure reagents used with resin sections Many of the staining reagents in routine use, such as enzyme substrates, dyes and fluorochromes, are impure. Industrial syntheses typically give rise to mixtures of homologues and isomers. Staining solutions may thus be mixtures of compounds of quite different log P values. Therefore, one batch of stain may give no background, whilst the next batch gives artefactual staining. One example seen by us involved the disulphonated acid dye Azo Fuchsin B. Chromatography showed that the batch giving background staining contained a substantial amount of a more lipophilic monosulphonated contaminant. Another example was provided by the observation that Methyl Green-Pyronin staining gave background staining with some dye batches, but not with others. I n this context, the contamination of many commercial samples of the hydrophilic dye Methyl Green by lipophilic Crystal Violet (e.g. Bancroft & Stevens, 1982) is of interest. Changes in lipophilicity can also arise due to decomposition of a reagent in the staining solution. This appears to have occurred with the tetra-cationic dye Cuprolinic Blue (see the Results section). T h e increased background staining seen when using alkaline staining solutions may be attributed to hydrolytic conversion of one or more hydrophilic quaternary ammonium cations into non-ionic lipophilic bases. Choosing optimum staining reagents-hydrophobicity and size effects It is now possible, when selecting staining reagents for GMA sections, to consider both hydrophobic and size effects. A numerical indication of such influences may be given diagrammatically, as in Fig. 3.
Lipophilic reagents and water-miscible resins
205
CONCLUSIONS
(1) Stain-resin hydrophobic effects can result in both understaining of tissue and overstaining of G M A resin. ( 2 ) Some staining methods developed for cryostat and paraffin sections will suffer from such artefacts when applied to GMA sections. ( 3 ) To avoid hydrophobic and size artefacts, the numerical structure-activity model summarized in Fig. 3 is useful. (4) I n particular, hydrophobic and size artefacts are least likely with reagents having negative log P values and ionic/molecular weights of
Staining sections of water-miscible resins
2. Effects of staining-reagent lipophilicity on the staining of glycol-methacrylate-embedded tissues by R I C H A R D w. HOROBIN*, P E T E R 0. G E R R I T S t and D A V I D J. WRIGHT*', *Department of Biomedical Science, The University, Shefield S10 2 T N , U.K. and tAnatomy and Embryology Laboratory, The University of Groningen, Oostersingel69, 97 13 E Z Groningen, The Netherlands
w O R D S . Artefact, glycol methacrylate (GMA), lipophilicity, structure-staining correlations, structure-activity relationship, water-miscible resin.
KEY
SUMMARY
Glycol methacrylate (GMA) sections of animal tissues were stained with a group of twenty-seven reagents of very varied chemical characteristics. T h e artefactual background staining of the resin was found to be dependent on the hydrophilic/lipophilic character of the staining reagent, as estimated from the logarithm of its octanol-water partition coefficient (log P). Intense background staining occurred with lipophilic stains, whose log P > 2 . I n keeping with this, use of G M A semi-permeable membranes for enzyme histochemistry failed to give staining when using a lipophilic substrate, probably because the substrate was trapped in the membrane. An analysis of other routine histochemical stains-in terms of the probable occurrence of high resin background staining and low tissue sensitivity-is made. A numerical guide is provided to help avoid artefacts resulting from hydrophobic and size effects. Note: small, hydrophilic reagents (log P < 0; molecular weight < 550 Da) are least likely to show either type of artefact. Conversely, reagents which are lipophilic, or/and of intermediate size (log P z 2; 550 < ionic weight < 1000 Da), give strong background staining. INTRODUCTION
T h e use of resin-embedded sections, especially of glycol methacrylate (GMA), is now commonplace in light microscopy. T h e advantages of using GMA include ease of cutting thin sections, even of hard tissues, and the applicability of staining procedures developed for paraffin and cryostat sections. However, such staining methods may fail with GMA sections, and may give high background staining of the resin itself. We have sought to clarify such problems, and were initially concerned with the influences of reagent molecular size (Gerrits et al., 1990). I n this earlier study all the reagents used were hydrophilic, since a pilot investigation indicated that lipophilic dyes gave high resin background staining. It is such background lipophilia which are the subject of this present paper.
'
Present address. Department of Zoology, 1 1 17 West Johnson, The University of Wisconsin, Madison, WI 53706, USA. (
1992 The Royal Microscopical Society
199
200
R. W . Horobin, P. 0. G e r r i t s and D . J . Wright
Since we wished to evaluate the role of stain-resin hydrophobic bonding, it was necessary to exclude other affinity or rate effects. Artefactual staining of the resin by basic dyes was avoided by using methacrylic acid-free resin (see Gerrits & van Leeuwen, 1987). In choosing staining reagents, effects due to size-dependent rate of penetration had to be avoided. Consequently, most reagents used in the present study are small, with ionic weights < 550 Da. However, reagents of moderate size can give useful information when they fail to stain the resin, although with such reagents resin staining is ambiguous, since it could be due to the effects of the rate of stain penetration (Gerrits et al., 1990). With these issues in mind, stain-resin hydrophobic bonding was investigated using staining reagents which varied from very hydrophilic to markedly lipophilic. It was supposed that background staining of the resin due to hydrophobic bonding would be most intense with lipophilic compounds. T h e hydrophilicity-lipophilicity of the reagents was modelled numerically using the log P parameter, i.e. the logarithm of the octanol-water partition coefficient (Hansch & Leo, 1979). MATERIALS AND METHODS
Preparation of resin sections for staining with dyes Various rat tissues were fixed for 24 h in 4O, neutral buffered aqueous formaldehyde. Tissues were dehydrated through the ethanols, and embedded in G M A using the JB4 kit (Polyscience, Warrington, U.S.A.), according to their Datasheet No. 123. Sections were cut at 2 pm with a glass knife, on an Anglia Scientific AS 500 microtome. Sections were floated onto water at room temperature, mounted on slides, and dried down at 37°C for 2 h or at 60-C for 10 min. Preparation of resin sections for enzyme histochemistry This was carried out in a similar way, as described in detail in Gerrits et al. (1989). This involved covering the resin tissue with a 2-pm section of tissue-free GMA, as a semi-permeable membrane. T h e dyes used are listed in Table 1, together with their sources and certain physicochemical characteristics. T h e log P values were obtained by calculation; the procedure is described in detail in Hansch & Leo (1979). T h e purity of the dyes was assessed using the reverse-phase thin-layer chromatography system of Proctor & Horobin (1985). Staining of resin sections Staining solutions were made up by dissolving 0.1 g of dye in 100 ml of aqueous buffer. T h e buffers used were 0.1 M p H 4 acetate, 0.1 M p H 7 phosphate, and 0.1 M p H 9 barbitone. Non-ionic dyes were used at p H 7; most acid and basic dyes were used at both p H 4 and 9. T h e only exception to this was Aldehyde Fuchsin, which was prepared following the procedure of Wolters et al. (1983), and was thus dissolved in acid alcohol usual for this reagent. Sections were stained until no further increases in coloration were seen, washed in running water for 1 min, and blotted or blow dried with compressed air. Sections were then air dried for 20 min, dipped in xylene, and mounted in a synthetic resin (DPX). E n z y m e histochemistry on G M A sections Non-specific esterase (i.e. 1-naphthyl acetate esterase) was demonstrated using diazotized Pararosanilin as the coupler (Bancroft & Stevens, 1982). Alkaline phosphatase was demonstrated using Naphthol AS-MX phosphate as the substrate and Fast Blue B as the visualizing agent (Gerrits et aZ., 1989). Acid phosphatase was demonstrated using 1-naphthyl phosphate as the substrate, and diazotized Pararosanilin as the coupler (after Barka & Anderson, 1963).
Lipophilic reagents and water-miscible resins
20 1
Table 1. Reagents used for staining GMA sections.
Reagent (Colour Index identification number) Acetoquinone Blue BF (61545) Astrazone Orange R (48040) Azo Fuchsin B (16550) Biebrich Scarlet (26905) Bismark Brown R (21010) Brazilein (-) Chromotrope 2R (16570) Crystal Violet (42555) Cuprolinic Blue (-) Dispersol Blue 7G (62500) Eosin Y (45380) Ethyl Violet (42600) Fast Blue BB (37175) Hematein (-) Hexazotized Pararosanilin (-) Indanthrene Bordeaux B (60705) Janus Green B (1 1050) Mordant Brown 6 ( 11875) Naphthol AS-BI phosphate (-) Naphthol AS-MX phosphate (-) 1-Naphthyl acetate (-) 1-Naphthyl phosphate (-) Orange G (16230) Purpurin (75410) Rhodamine 6G (45 160) Thionin (52000) Toluidine Blue (52040)
log P*
+ 0.2 + 2.2 - 5.0
- 3.2 - 0.4
+ 1.3 - 4.3 + 1.2 - 14.5 + 1.3 + 1.4 + 3.4 - 0.4 + 1.2 < -5 + 0.9 + 0.4 - 0.4 + 0.3 i 1.4 + 3.4 - 1.0 - 5.0
+
+ 1.0 3.5
- 2.1 - 0.6
Ionic or molecular weight (Da) 326 39 1 436 510 390 286 422 372 640 358 646 456 312 300 312 498 475 316 294 369 167 222 406 256 443 225 270
Charge 0 +1 -2 -2 +l 0
-2 +1 +4 0 -2 +1 +1 0 +3 0 +1 -1 -1 -1 0 -1 -2 0 +1 +1 +1
Stain sourcet EG EG EG EG AL EG EG EG BD MF AL KL SI KL AL EG BD AL SI SI SI SI AL EG AL AL AL
The log of the octanol-water partition coefficient. BD = British Drug Houses, E G = Edward Gurr, K L = Koch Light, M F = manufacturer's sample.
t AL = Aldrich,
Assessment of relative staining intensities This was carried out subjectively by three independent observers. RESULTS
Overall summary of background resin staining Background staining of resin in sections may be summarized as follows. (1) T h e following dyes gave no staining of the resin: Azo Fuchsin B (sample l), Chromotrope 2R, Cuprolinic Blue (acidic solution), Orange G, Thionin. (2) T h e following dyes gave a trace of resin staining when viewed macroscopically, but this was not noticed in the microscope: Azo Fuchsin B (sample 2 ) , Biebrich Scarlet, Bismark Brown, Cuprolinic Blue (alkaline solution), Mordant Brown 6, Toluidine Blue. ( 3 ) T h e following dyes stained the resin weakly but sufficiently to be noticed in the microscope: Acetoquinone Blue BF, Brazilein, Dispersol Blue 7G, Eosin Y, Hematein, Indanthrene Bordeaux B, Purpurin. (4) T h e following dyes stained the resin intensely, obscuring the specimen when viewed microscopically: Astrazone Orange R, Crystal Violet, Ethyl Violet, Janus Green B, Rhodamine 6G.
202
R. W . Horobin, P. 0. Gerrits and D . J . Wright
% 0 0
8 w
0 I
none
trace
weak
intense
Degree of resin background staining Fig. 1. A summary of the influence of the hydrophilic lipophilic character of stains on the degree of resin background staining.
Fig. 2. An illustration of the effect of reagent lipophilicity on resin background staining. Section (a) was stained with the hydrophilic basic dye Thionin, whilst section (b) was stained with the lipophilic basic dye Janus Green B. Arrows indicate regions of tissue-free resin: black arrows, unstained resin; white arrows, stained resin. Magnification x 550.
Lipophilic reagents and water-miscible resins
203
Enzyme histochemical staining. Acid and alkaline phosphatases were successfully demonstrated using the GMA-membrane technique, whereas non-specific esterase was not. Influences of hydrophobic effects on the background resin staining of sections As expected (Tanford, 1980), hydrophobic effects increased as the lipophilicity of the staining reagents rose, as seen in Fig. 1. Since, for practical purposes, background staining not noticeable in the microscope may be ignored, Fig. 1 suggests the following rules of thumb. Small molecular weight dyes with log P < 0 will usually give minimal or no hydrophobic background staining; dyes with log P > 2 will often give gross background coloration of resin; dyes of intermediate lipophilicity give weak staining of the resin. These conclusions are illustrated by the micrographs seen in Fig. 2. Background staining can of course occur for reasons other than hydrophobic effects, as described in the Introduction. Reagent-resin absorption due to hydrophobic effects will still occur with colourless staining reagents. However, the observed outcomes are now likely to take the form of understaining of the tissue, rather than overstaining of the resin. An example of this was provided by comparing our successful demonstration of acid and alkaline phosphatases with our failure to demonstrate esterase. All the enzyme substrates and visualizing agents used were small, with ionic weights < 500 Da, so staining failure would not have been a size effect (cf. Gerrits et al., 1990). However, log P values of the enzyme substrates differed markedly. T h e ineffective substrate was by far the most lipophilic (log P = 3.4), whilst effective substrates had log P values of - 1.0 and 1.4. It may be supposed that the amount of the more lipophilic substrate reaching the tissue was markedly reduced due to binding to the resin of the overlying membrane. T h e log P cut-off point for effective enzymatic staining is consistent with the results seen with the dyes.
+
+
DISCUSSION
Implications of hydrophobic effects for everyday microscopy Excessive background staining of resin, or understaining of tissue, due to reagent lipophilicity has marked implications for the everyday use of G M A sections. We must ask: what routine staining reagents might give staining artefacts due to the hydrophobic effect? This may be put numerically, since the rule of thumb given above suggests that significant problems may occur with reagents having log P values > 1. Considering dyes and fluorochromes, many anionic compounds are hydrophilic. However, the classic acid dye Eosin Y is slightly lipophilic, and gives sufficient background staining to require an alcohol rinse. T h e commonly used basic dyes and cationic fluorochromes are more often lipophilic. Thus background staining is substantial with Aldehyde Fuchsin and Crystal Violet, and may be expected with reagents such as Granular Blue or Pyronin B. T h e same is true for non-ionic dyes and probes, e.g. Nile Red or the Sudan dyes. However, it is sometimes possible for such reagents to stain lipids in GMA sections, by use of careful differentiation in a plasticizing solvent (cf. Gerrits et al., 1987). T h e most commonly used reactive reagent, Schiff’s, is hydrophilic and gives no problems. T h e same applies to the calcium complexing agent Alizarin Red. However, the reactive thiol reagent Mercury Orange is lipophilic and can give high resin backgrounds, whose removal requires extraction with plasticizing solvents. Routinely used enzyme substrates include several which are lipophilic, with log P values > 1. These include certain naphthyl and indoxyl compounds and their halogenated derivatives, e.g. those used for cholinesterase, esterase and lipase
204
R. W . Horobin, P. 0. Gerrits and D . J . W r i g h t
this zone was not investigated
0
logP
2 r*
Reagent hydrophilicity/lipophilicity Fig. 3. A guide for choosing artefact-free staining reagents for GMA sections. Reagents falling into zone 1 will give few problems. Those in zone 2 will give some resin background, removable with a plasticizing solvent such as ethanol. Reagents falling into zone 3 give no background, but only stain resin-free or surfaceexposed tissue components. Reagents in zone 4 give intense resin background staining.
demonstrations. Peroxidase substrates, e.g. naphthols and arylamines, are also generally lipophilic. Such compounds are used both in enzyme histochemistry and in enzyme-labelled immunostaining.
Problems due to impure reagents used with resin sections Many of the staining reagents in routine use, such as enzyme substrates, dyes and fluorochromes, are impure. Industrial syntheses typically give rise to mixtures of homologues and isomers. Staining solutions may thus be mixtures of compounds of quite different log P values. Therefore, one batch of stain may give no background, whilst the next batch gives artefactual staining. One example seen by us involved the disulphonated acid dye Azo Fuchsin B. Chromatography showed that the batch giving background staining contained a substantial amount of a more lipophilic monosulphonated contaminant. Another example was provided by the observation that Methyl Green-Pyronin staining gave background staining with some dye batches, but not with others. I n this context, the contamination of many commercial samples of the hydrophilic dye Methyl Green by lipophilic Crystal Violet (e.g. Bancroft & Stevens, 1982) is of interest. Changes in lipophilicity can also arise due to decomposition of a reagent in the staining solution. This appears to have occurred with the tetra-cationic dye Cuprolinic Blue (see the Results section). T h e increased background staining seen when using alkaline staining solutions may be attributed to hydrolytic conversion of one or more hydrophilic quaternary ammonium cations into non-ionic lipophilic bases. Choosing optimum staining reagents-hydrophobicity and size effects It is now possible, when selecting staining reagents for GMA sections, to consider both hydrophobic and size effects. A numerical indication of such influences may be given diagrammatically, as in Fig. 3.
Lipophilic reagents and water-miscible resins
205
CONCLUSIONS
(1) Stain-resin hydrophobic effects can result in both understaining of tissue and overstaining of G M A resin. ( 2 ) Some staining methods developed for cryostat and paraffin sections will suffer from such artefacts when applied to GMA sections. ( 3 ) To avoid hydrophobic and size artefacts, the numerical structure-activity model summarized in Fig. 3 is useful. (4) I n particular, hydrophobic and size artefacts are least likely with reagents having negative log P values and ionic/molecular weights of
