Fluorescence diagnosis of bladder cancer: a novel in vivo approach using 5-aminolevulinic acid (ALA) dendrimers

Abstract

What's known on the subject? and What does the study add?

Fluorescence cystoscopy with hexylaminolevulinate (h-ALA, Hexvix®) is known to improve tumour detection in non-muscle-invasive bladder cancer. However, specificity is relatively low and the intensity of the observed fluorescence signal decreases over time due to protoporphyrin IX (PpIX) efflux.

This study evaluates in an in vivo model the use of a dendritic 5-aminolevulinic acid compound for fluorescence diagnosis. Fluorescence ratios between tumour and urothelium as well as muscle were significantly better as compared with h-ALA. Sustained synthesis of PpIX accounts for preservation of fluorescence for >24 h.

OBJECTIVE

• • 
  To overcome the relative lack of tumour selectivity of fluorescence-guided cystoscopy using 5-aminolevulinic acid (ALA) or its ester derivative (e.g. hexylaminolevulinate, h-ALA; Hexvix®), we evaluated the use of dendrimers bearing different ALA loads in rats bearing orthotopic bladder tumours.

MATERIALS AND METHODS

• • 
  Rat bladders were instilled with h-ALA or ALA dendrimers and fluorescence ratio between tumour and normal urothelium, as well as tumour and muscle and depth of fluorescence were determined with Image J software.
• • 
  Quantification of ALA and/or esters systemic reabsorption was evaluated by high-performance liquid chromatography.

RESULTS

• • 
  Slow hydrolysis of ALA from dendrimers as observed in vitro implies a higher initial ALA load and longer resting times in vivo. Sustained synthesis of protoporphyrin IX (PpIX) explains persistence of fluorescence for >24 h.
• • 
  There were significantly better fluorescence ratios with dendrimers, as well as higher penetration depths and absence of systemic reabsorption.

CONCLUSION

• • 
  The prolonged and sustained PpIX synthesis, the improved tumour selectivity with a deeper penetration and the absence of systemic reabsorption are primary indicators that ALA dendrimers could be an alternative to h-ALA in fluorescence-guided cystoscopy.

Abbreviations

ALA

5-aminolevulinic acid

FCS

fetal calf serum

h-ALA

hexylaminolevulinate

PBG

porphobilinogen

PpIX

protoporphyrin IX

INTRODUCTION

Fluorescence cystoscopy is based on the intravesical administration of a drug, which will be incorporated to a higher degree into tumour cells and result in fluorescence, revealing the presence of cancerous lesions. The concept was first published in 1996, by Kriegmair et al. [1] who described a sensitivity of 97% after 5-aminolevulinic acid (ALA) administration. ALA is taken up by the cells and metabolized through the heme cycle into a fluorescent compound, protoporphyrin IX (PpIX). Shortly thereafter Lange et al. [2] showed that the use of an ALA ester (hexylaminolevulinate, h-ALA, Hexvix®), induced a more intense fluorescence with reduced drug concentration. As multiple large studies have reported the major benefits that can be obtained through fluorescence-guided resection, irrespective of the prodrug used [3,4] without differences between 5-ALA and Hexvix [4]. An enhanced sensitivity of detection (85–95%), leads to a more complete resection, resulting in a reduced recurrence rate (90% tumour free at 1 year vs 70% with white light). Recently, a prospective randomised international study of 814 patients also showed a significantly improved detection rate [5]. However, in two multicentre double-blind randomised placebo-controlled trials, the authors failed to recognise any long-term benefit in terms of recurrence [6,7]. Moreover, major drawbacks, are the lack of tumour selectivity, the very superficial fluorescence and the efflux of PpIX a few hours after instillation [8,9]. These limitations can be reduced by developments in optical techniques, e.g. narrow-band imaging, Raman spectroscopy and optical coherence tomography, but necessitating costly equipment [10]. There is thus a real need for other fluorescent markers that are more specific and less prone to photodegradation.

A current focus of pharmaceutical research is the development of nanocarrier systems to enhance drug delivery and specificity. Dendrimers are hyperbranched polymers that have attracted considerable interest in oncology for drug delivery, owing to their ability to carry high drug-payloads and their well-defined macromolecular structure [11]. Recently, a dendrimer containing 18 ALA molecules coupled through biodegradable ester linkages was synthesised and tested in vitro and in vivo[12,13]. Those studies indicated a more efficient and sustained PpIX production than ALA at drug equivalent doses after i.p. administration.

The aim of the present study was to compare the in vitro behaviour of ALA, h-ALA (Hexvix) and ALA dendrimers in a bladder cancer cell line. The second objective was to compare the impact of intravesical administration of ALA dendrimers and h-ALA (Hexvix) on tumour specificity and depth of penetration of PpIX in rats bearing orthotopic bladder tumours.

MATERIALS AND METHODS

BLADDER CANCER CELL LINE, ALA-DERIVATIVES PRODRUGS AND ANIMAL MODEL

AY27 cells were initially derived from carcinomas of the urinary rat bladder chemically induced with N-(4–{5-nitro-2-furyl}–2-thiazolyl) formamide (FANFT). Cells were cultured in vitro as a monolayer in RPMI 1640 medium (Gibco, Cergy Pontoise, France) supplemented with 9% fetal calf serum (FCS; PAN Biotech GmbH, Aidenbach, Germany), 1% l-glutamin (200 mm) and 1% penicillin (10 000 UI), streptomycin (10000 µg/mL) (Gibco).

The prodrugs used were 5-ALA (Sigma-Aldrich, France), hexylaminolevulinate (h-ALA; Photocure, Oslo) and ALA dendrimers (6-ALA and 18-ALA; Organix, Essex, UK) and were freshly prepared in PBS for in vivo experiments or in serum-free medium for in vitro experiments, as serum is known to cause release of PpIX from the cells, thus resulting in loss of the fluorescence signal. Addition of prodrugs at non-toxic concentrations did not change the pH of the cell medium. ALA dendrimer (6-ALA dendrimer) is a first-generation dendrimer, bearing six ALA molecules with free amino groups, while 18-ALA dendrimer is a second-generation dendrimer, bearing 18 ALA molecules with free amino groups (Fig. 1A,B). The ALA moieties are coupled via ester bounds so that ALA can be released via non-specific esterases.

Figure 1. Structure of A: 6-ALA dendrimer, B: 18-ALA dendrimer.

IN VITRO FLUORESCENCE PHARMACOKINETICS OF PPIX, DETERMINATION OF INTRACELLULAR ALA CONTENT AND ALA HYDROLYSIS

All experiments were performed at least in triplicate and under subdued light. After removing the culture medium, exponential growing cells were incubated for 2 h with freshly prepared solutions of ALA or its derivatives. Controls received serum-free medium without ALA or derivatives but were otherwise treated in the same way as the other samples. After incubation, incubation solutions were removed and replaced with 2% FCS medium for different times before analysis.

FLUORESCENCE PHARMACOKINETICS

After incubation with the prodrugs, cells were trypsinized and both flow cytometry (FACScalibur cytometer, Becton-Dickinson, Heidelberg, Germany) and fluorescence spectroscopy were used to collect the PpIX fluorescence signal. For spectroscopic measurements, PpIX was extracted from cells by ice-cold lysis extraction mixture, consisting of ethanol/dimethyl sulphoxide/acetic acid (80/20/1, v/v/v). The cell lysate was centrifuged (400 g, 10 min, 2 °C), the pellet was grinded and the suspension was sonicated for 15 min before centrifugation. The fluorescence signal of PpIX was measured with Perkin-Elmer LS 55 spectrometer (Perkin-Elmer, Beaconsfield, UK) and expressed in function of the protein content.

INTRACELLULAR ALA CONCENTRATION AND ALA ESTERS HYDROLYSIS

After incubation with h-ALA or 18-ALA for 2 h, cells were lysed with 5% trichloroacetic acid. Concentrations of ALA and porphobilinogen (PBG), were determined according to the method described by Perotti et al. [14]. 5-ALA and PBG are known to be early intermediates in the biosynthesis of porphyrins. For ALA measurement, a condensation reaction at 100 °C in presence of acetylacetone was perfomed to induce the formation of pyrrole rings that were quantified after centrifugation by adding Ehrlich reagent (2% dimethylaminobenzaldehyde in 6 m HCl). For PBG determination, Ehrlich reagent was added directly to supernatant without condensation. ALA values were obtained by subtracting PBG values from the total condensed pyrroles.

Quantification of ALA esters was carried out by ion-exchange chromatography, as previously reported [15]. ALA was separated from ALA derivatives, and the percentage of the total was calculated from the total ALA after subtracting the PBG value.

ANIMAL TUMOUR MODEL AND PRODRUG ADMINISTRATION

All animal procedures were performed according to institutional and national guidelines defined by the French ministry of agriculture N°2001-464. Female Fischer (F344) rats were purchased from HARLAN Laboratories (Gannat, France). The rats were anaesthetised with i.p. sodium pentobarbital injection (45 mg/kg; Ceva Sante Animale, Libourne, France). The rat bladder was catheterised via the urethra with a 14 G i.v. cannula (Terumo Surflo, Guyancourt, France) and drained. The orthotopic bladder cancer model, previously described by Xiao et al. [16], was based on an epithelial desquamation by means of an intravesical instillation of 0.5 mL chlorhydric acid (HCl, 0.1 m) for 15 s, neutralised by adding 0.5 mL of sodium hydroxide (NaOH, 0.1 m). The bladder was then drained and flushed three times with PBS. Immediately after bladder conditioning, freshly harvested AY27 cells (10⁶ cells in 0.5 mL medium) were instilled intravesically for 1 h. Experiments were carried out 14 days after implantation of tumours. Rats (n= 35) bearing orthotopic bladder tumours were instilled with 0.5 mL of the various prodrugs for 1 or 2 h, with different resting times to determine optimal conditions with a maximum fluorescence ratio of tumour/normal tissue.

FLUORESCENCE MICROSCOPY

Before cystectomy, bladders were filled with tissue-freezing medium, and snap-frozen in isopentane precooled in liquid nitrogen. Sections of 5 µm were observed with an epifluorescence microscope (AX-70 Provis Olympus, Rungis, France). Image J software (National Institute of Health, Bethesda, MD, USA) was used to quantify fluorescence from normal epithelial, tumour and muscle for the different instillation conditions. Two consecutive slices were used for histology with haematoxylin eosine safran.

QUANTIFICATION OF SERUM ALA AND ESTERS CONTENT IN PLASMA BY HPLC

Rats were separated in the following groups: healthy bladder controls, healthy bladders instilled with prodrugs, bladder tumours control and bladder tumours instilled with prodrugs. After different resting times, blood samples were collected by heart puncture in heparinised tubes.

ALA and esters quantification in blood samples was based on the conversion of ALA or ALA esters into a fluorescent derivative by a condensation reaction according to the Hantzsch reaction. Briefly, as previously described [17], plasma were mixed with 25 µL HCL 0.1 m and 1 mL acetylacetone reagent consisting of acetylacetone (Sigma-Aldrich, France), ethanol 95% (Elvetec, France), deionised water (15/30/55, v/v/v) and 100 µL formaldehyde 10% (Sigma-Aldrich). Samples were then heated for 40 min at 60 °C and immediately cooled down in ice for 10 min before centrifuging at 2140 g for 5 min. The chromatographic system was composed of a programmable solvent module (System Gold 126, Beckman Coulter, Fullerton, CA, USA), an autosampler injector (507e, System Gold, Beckman Coulter), and a scanning fluorescence detector (RF-10A XL, Shimadzu, Kyoto, Japan). Analyses were performed by reverse-phase HPLC on a C18 analytical column (250 × 4.6 mm internal diameter, S-5 µm, YMC, Interchim, Montluçon, France), under isocratic elution conditions at 40 °C with a mobile phase consisting of methanol/deionised water/acetic acid (70/30/1) at a flow rate of 1 mL/min. The retention time of ALA residues was 4.9 ± 0.2 min. Quantifications were based on peak areas of the chromatogram, acquired and analysed using GOLD Nouveau software version 1.6 (Beckman Coulter) and deduced from calibration curves.

PLASMA PPIX MEASUREMENT

The extraction of PpIX from plasma was based on the precipitation of the plasma proteins by acetonitrile, as previously described by Collaud et al. [17]. After sonication and centrifugation (4500 g), a fraction of the supernatant was mixed with acetate buffer Ph 5.2 (v.v⁻¹). The fluorescence intensity was monitored with spectrofluorimeter by adding a volume of known PpIX concentrations to the extracted plasma sample. The validated detection range of PpIX in plasma was 10–100 ng/mL.

STATISTICAL ANALYSIS

The P value was determined with Mann–Whitney test (Statview software) and was considered significant when P < 0.05.

RESULTS

PPIX FLUORESCENCE PHARMACOKINETICS AND ALA HYDROLYSIS

We first established the appropriate concentration of 18-ALA to obtain in vitro an identical PpIX fluorescence signal as compared with h-ALA (Fig. 2). We have previously established that 0.8 mm h-ALA concentration was the highest non-toxic concentration for AY27 cells incubated for 2 h [18]. It appears that only ALA esters induce a prolonged PpIX production, whereas PpIX fluorescence induced by ALA drops dramatically after removal of ALA (Fig. 2). Incubation with dendrimers at the lowest concentrations (0.05, 0.1 mm) results in a progressive decrease in PpIX fluorescence after removal of the drug. Only the highest concentration (0.2 mm) induces a comparable fluorescence signal to that seen using h-ALA 0.8 mm. Both concentrations have previously been shown to have minimal dark toxicity [12,18]. Since the ALA pay-load in dendrimers is much higher, the amount of ALA that is provided can be expressed in molar equivalents of ALA as shown in Table 1. It then appears that the amount of ALA needed to produce comparable fluorescence intensities is four-times higher in the case of 18-ALA (3.6 mm ALA) as compared with h-ALA (0.8 mm ALA).

Figure 2. Fluorescence pharmacokinetics of PpIX. The mean PpIX fluorescence emission intensity (in arbitrary units) from AY27 cells induced by ALA derivatives, assessed by flow cytometry. Cells were incubated with prodrugs for 2 h. Results are presented as the mean ±sd of at least four experiments.

Table 1. ALA equivalent dose (amount of ALA, as expressed in the ALA molar equivalent value) according to prodrug concentration

Prodrugs	[Prodrugs], mm	ALA molar equivalent, mm
h-ALA	0.8	0.8
18-ALA	0.05	0.9
	0.1	1.8
	0.2	3.6
	0.5	9
	0.7	12.6

As measured by chemical extraction, 24 h after the end of the incubation with 18-ALA, PpIX production is significantly higher (P < 0.05) in comparison with h-ALA and can still be detected 3 days later (Fig. 3). Prolonged PpIX synthesis, as well as the need for higher initial ALA load can be explained by the slow hydrolysis of ALA from dendrimers (Fig. 4). Hydrolysis of ALA from the dendritic structure after incubation in AY27 cells is slow, as compared with h-ALA, with 50% free ALA at 7 h for h-ALA as compared with 15% when dealing with dendrimers (Fig. 4). Hydrolysis from dendrimers follows a mono-exponential decay (k 6.87 h⁻¹) whereas esterase activity is faster in case of h-ALA and is bi-exponential (k₁ 32.06 h⁻¹, k₂ 3.88 h⁻¹). Moreover, hydrolysis from dendrimers remains stable for at least 24 h, which results in a sustained PpIX production as shown in Fig. 3.

Figure 3. PpIX extraction from ALA-derivatives in AY27. Cells were incubated with 0.8 mm h-ALA or 0.2 mm 18-ALA dendrimers for 2 h. Each data point represents the average of at least four experiments. *P < 0.05

Figure 4. ALA hydrolysis in AY27 cells after 2 h incubation with 0.8 mm h-ALA or 0.2 mm 18-ALA. Results are presented as the mean ±sd of three experiments. Regression coefficient (r²) for h-ALA 0.9932; r² for 18-ALA 0.9944.

TUMOUR SPECIFICITY ACCORDING TO INSTILLATION PROTOCOLS

Optimal in vivo instillation conditions for h-ALA were previously reported by our group [19]. h-ALA at 8 mm has to be instilled for 1 h followed by 2 h resting time to induce a maximal PpIX fluorescence in the tumour. Slow hydrolysis has to be taken into account when optimising in vivo parameters for 18-ALA dendrimers. Indeed, the resting time to build up detectable PpIX fluorescence after 18-ALA instillation has to be increased from 2 to 3–4 h (Table 2, Fig. 5). With this time regime and using identical ALA molar equivalents (8 mm ALA), the tumour/muscle fluorescence ratio was identical to h-ALA (ratio ≈2.5) but the fluorescence ratio of tumour to normal urothelium was increased (1.7 vs 1) (Table 2). Considering the slow hydrolysis of dendrimers, we increased the ALA molar equivalent by 50% (12.6 vs 8 mm).

Table 2. In vivo fluorescence analysis. Comparison of tumour specificity according to different instillation protocols. Fluorescence intensities were determined using ImageJ software. Fluorescence ratios were calculated on three rat bladders for each experiment

[Prodrug]	ALA equivalent dose, mm	Molecular weight, g.mol^−1	Instillation time, h	Resting time after instillation, h	Fluorescence ratio, mean (sd)		Depth penetration, µm
					tumour-to-muscle	tumour-to-healthy urothelium	Mean (sd)	Maximum
*P < 0.05; †P < 0.01 vs h-ALA.
h-ALA, 8 mm	8	251.8	1	2	2.5 (0.8)	1.2 (0.4)	115 (60)	250
18-ALA, 0.5 mm	9	4317.5	1	4	2.4 (1.1)	1.7 (0.2)	108 (35)	400
18-ALA, 0.7 mm	12.6		1/2	4	3.4 (1.1)†	2.9 (1.1)†	247 (66)*	750
			1/2	24	2.3 (0.9)	2.0 (0.6)	189 (72)	550
6-ALA, 1.5 mm	9	1716	1/2	2/3	2.0 (0.7)	1.2 (0.4)	121 (40)	500

Figure 5. Morphology and fluorescence microscopy of rat bladder tumours after instillation of a) h-ALA 8 mm (1 h + 2 h) and b) 18-ALA 0.7 mm (1 h + 4 h). ur, normal urothelium; M, muscle; T, tumour.

When applying those parameters, there was a significant (P < 0.01) increase in the tumour-to-muscle fluorescence ratio (from 2.5 to 3.4) as well as tumour to normal urothelium (from 1.2 to 2.9) as compared with h-ALA (Table 2). Fluorescence is still seen 24 h after 18-ALA instillation (Table 2), whereas fluorescence completely disappears 4 h after h-ALA [19]. PpIX fluorescence can be seen in the deeper tumour layers up to 120 µm for h-ALA whereas a mean and significant (P < 0.05) fluorescence depth of 250 µm for 18-ALA (Table 2, Fig. 5).

We assumed the enhanced specificity of the fluorescence signal, especially towards normal urothelium, to be due to the size/molecular weight of the different compounds. We therefore tested smaller dendrimers with a load of six ALA molecules and applied an identical time regime (1 h instillation, 3 h resting time) and ALA molar equivalents (8–9 mm; Table 2). Considering tumour to normal urothelium ratios there was no significant difference between h-ALA and 6-ALA instillation (1.2 vs 1.2). However, 18-ALA dendrimers had a significant better ratio of 1.7. The depth of penetration of PpIX signal into the tumours did not vary according to the prodrug.

HPLC ASSESSMENT OF PLASMA REABSORPTION OF ALA RESIDUES

To assess potential systemic toxicity after intravesical instillation, we assessed the presence of free ALA, ALA esters and PpIX in the plasma after instillation of the three compounds. Plasma samples failed to show the presence of PpIX (data not shown). ALA and their esters are found at the highest concentrations for h-ALA, but were significantly (P < 0.05) lower for 6-ALA and could not be detected (P < 0.01) when using 18-ALA (Table 3). Observations for non-tumour bearing bladders are identical but at higher concentrations due to the absence of retention in the tumour.

Table 3. Concentrations of ALA / esters in plasma after bladder instillation of prodrug

[Prodrugs]	ALA equivalent dose, mm	Instillation time, h	Resting time after instillation, h	Mean (sd) ALA (free or esters), ng/mL
*P < 0.05; †P < 0.01 vs h-ALA. <d.l.^a, detection limit of ALA residues by HPLC ≤4 ng/mL.
Tumour-bearing rats:
No instillation	–	–	–	<d.l.^a
h-ALA, 8 mm	8	1	2	28.3 (7.9)
18-ALA, 0.7 mm	12.6	1/2	4/24	<d.l.^a*
6-ALA, 1.5 mm	9	1/2	2/3	4.3 (2.4)*
Healthy rats:
No instillation	–	–	–	<d.l.^a
h-ALA, 8 mm	8	1	2	79.2 (9.4)
18-ALA, 0.7 mm	12.6	1/2	4/24	<d.l.^a†
6-ALA, 1.5 mm	9	1/2	2/3	38.1 (10.2)*

DISCUSSION

Fluorescence cystoscopy guided by ALA has been shown to benefit patients and provide more life years, but the high false-positive rates remains a major drawback [8]. This relative low specificity can be reduced when applying hexylaminolevulinate [2]. In the present study, we investigated the possibility of using a macromolecular delivery system based on a dendrimer. We have investigated two dendritic ALA derivatives: the 6-ALA dendrimer is a first-generation dendrimer, bearing six molecules of ALA, conjugated to the building blocks through esters bonds. The ester bonds were expected to be hydrolysed by non-specific esterase enzymes to yield the free ALA in the tissue. 18-ALA dendrimer is a second-generation dendrimer, carrying 18 ALA molecules (Fig. 1A,B). The exterior of the dendrimer is relatively hydrophilic at physiological pH due to protonation of the amino groups, whereas the interior structure is hydrophobic with an aromatic core and polyamido hydrocarbon spacers.

Previous studies showed that this dendrimer was incorporated in cells via endocytosis and resulted in an enhanced and sustained synthesis of PpIX in vitro as well as in vivo after i.p. administration [12,13]. However, specificity towards the tumour was essentially identical to that for ALA and never exceeded a factor of two [13]. Malik et al. [20] showed that similar non-ALA bearing dendrimers were rapidly cleared from the blood stream after i.p. or i.v. administration.

As a result, the amount of ALA delivered has to be increased when using dendrimers to reach an identical or enhanced PpIX production in vitro and in vivo (Fig. 3, Table 2). Slow hydrolysis of ALA from dendrimers (Fig. 4) also implies longer resting times to synthesise sufficient PpIX to visualize fluorescence in vivo (Table 2, Fig. 5). In our previous work, we showed that 1-h instillation with h-ALA in a rat orthotopic bladder tumour model followed by 2 h resting time provided an optimal PpIX fluorescence signal [19]. When dealing with 18-ALA the resting period must be extended to 3 h, which can be considered to be a clinical drawback. However, as the fluorescence persists at least as long as 24 h (Table 2), this might be very beneficial in a clinical setting. Indeed, the time between instillation and endoscopic inspection is no longer crucial and allows for more flexibility. It has indeed been shown that starting from 3 h after instillation, there is a progressive efflux of PpIX from cells, resulting in reduced fluorescence [21]. However, one has to consider that at 24 h, the specificity no longer exceeds the one observed with h-ALA. In a s.c. tumour model, Casas et al. [13] also reported sustained tumour fluorescence up to 24 h, progressively declining without however reaching basal levels at 48 h.

From an ultrastructural point of view, human and rat healthy bladder urothelium and carcinoma are comparable in many aspects such as for instance glycocalix layer and tight junctions [22,23]. It has been shown that changes occur in the junctional complex during carcinogenesis, resulting in enhanced permeability of the urothelium making it from tight to leaky. Furthermore, alterations in the glycocalix and the formation of microvilli on the luminal cell surface are seen, thus enhancing contact possibilities with intraluminal compounds. These conditions are partly responsible for the higher PpIX synthesis seen in bladder tumours. ALA and h-ALA, being small molecules with low molecular weight (Table 2), are capable of penetrating between normal urothelial cells, thus enhancing contact possibilities and internalisation, and potentially reducing tumour selectivity [24]. Small molecules have been shown to be more prone to enter the vasculature after intravesical administration [17] and to pass across fetal membranes after ex vivo application as opposed to dendrimers [25].

Thus, the size of an intravesically administered molecule appears to be essential to optimise tumour selectivity and depth of penetration into the tumour. Small molecules will tend to enter normal urothelium as well as the tumour, with loss of specificity. Penetration into the tumour depth would be facilitated but this would also potentially implicate enhanced penetration into the tumour vasculature. This would in turn result in reduced fluorescence in the deeper tumour layers as well as potential generalised toxicity due to systemic reabsorption, as previously shown [21].

We therefore investigated three compounds with different size/molecular weight; h-ALA 251.8 g/mol, 6-ALA 1716 g/mol and 18-ALA 4317.5 g/mol. At comparable ALA molar equivalents, tumour selectivity as compared with normal urothelium was the highest for 18-ALA and was identical for h-ALA or 6-ALA (1.7 vs 1.2, Table 2).

There were no differences between tumour/muscle ratios (±2.5) or depth of fluorescence visualisation (150 µm) at equimolar concentrations of 8–9 mm ALA When increasing the concentration, both parameters were significantly better for the largest molecules (Table 2). Tumour to muscle ratio increases to 3.4 with a mean penetration depth of 250 µm, reaching up to 750 µm. This can probably be attributed to the dendritic structure itself. Steric hindrance is responsible for the slow hydrolysis seen. However, by increasing the quantity of intracellular dendrimers, access to esterases is multiplied, resulting in enhanced PpIX production and fluorescence [26,27]. In case of h-ALA, increasing the concentration results in a saturation of the enzymatic pathway and reduced PpIX production as we have previously shown [28].

To further confirm the importance of size, we assessed free ALA, ALA esters and PpIX in plasma after instillation in normal and tumour-bearing bladders (Table 3). PpIX levels were always below detection limits of our set up (data not shown), although PpIX has been described in human plasma after h-ALA instillation [17]. ALA and esters were found in the highest quantities for h-ALA, less so for intermediate size dendrimers and were completely undetected for large size dendrimers. This observation also accounts for the deeper fluorescence seen after sensitisation with 18-ALA. Indeed, in the absence of systemic reabsorption, the dendrimers are in prolonged contact with the deeper tumour cell layers. Absence of systemic reabsorption of ALA is also an important factor when considering potential toxicity. Oral administration of ALA has been shown to induce temporary (<48 h) high plasma levels of ALA/PpIX, without a significant skin sensitisation effect but some cardiovascular effects, e.g. hypotension and tachycardia [29].

False-positive fluorescence signals emanating from non-transformed epithelium are a major drawback in fluorescence cystoscopy. This relative poor specificity decreases with increased experience with the technique as well as by using an esterified derivative. False-positive results have been attributed to several conditions, e.g. inflammation, hyperplasia or immunotherapy [1,30,31]. The fluorescence intensity from orthotopic rat bladder tumours never exceeds a factor of two compared with normal urothelium, irrespective of the prodrug used [19]. This can be attributed to the previous intravesical interventions for tumour induction. However, the present study achieves fluorescence ratios of three. This suggests that 18-ALA could be a valuable alternative prodrug to enhance sensitivity as well as specificity for tumour detection. Clinical trials are thus mandatory.

In conclusion, dendrimers bearing 18 ALA molecules can be instilled intravesically and induce sustained production of PpIX that can still be seen after 24 h. However, due to the slow hydrolysis of ALA, the resting time after instillation is longer (3–4 vs 2 h) after instillation. The concentration of the prodrug needed is significantly smaller than for 5-ALA (0.7 vs 180 mm) and h-ALA (0.7 vs 8 mm). Tumour specificity vs muscle and normal urothelium is significantly higher compared with h-ALA, the currently approved ALA derivative for fluorescence bladder cancer detection and the fluorescence signal can be seen at greater depths. This can be attributed to the molecular weight/size of the prodrug. There is no systemic reabsorption thus minimising generalised toxicity. 18-ALA dendrimers could be an alternative to h-ALA in fluorescence-guided cystoscopy.

ACKNOWLEDGEMENTS

We thank the French Ligue Contre le Cancer for financial support, Photocure ASA for providing Hexvix® and M. Barberi-Heyob for technical assistance with HPLC.

CONFLICT OF INTEREST

None declared.