Abstract One of the difficulties of porcine islet isolationis their fragility during collagenase digestion. The objectof this study was to determine the distribution of 4different collagen types within adult and juvenile porcinepancreata. Five different areas from each pancreas wereanalysed by light microscopy. The distribution of collagentypes I, IV, V and VI were measured within the interlobar,intralobular (acinar), peri-insular and intra-islet regions.Their was an abundance of collagen type VI compared toI, IV and V in both the interlobar and intralobular septa inboth juveniles (P<0.001) and adults (P<0.001). The peri-insular collagen content also showed diversity. This wasmainly attributable to the distribution of collagen type I(weak) and type VI (intense) in both adults and juveniles.In general, the peri-insular capsule was fragmentary andcontained less than 50% of the total islet circumference.The latter finding taken together with the distributions ofdifferent collagen types can partly explain some of thevariability of porcine islet isolation.

Key words Collagen · Pancreas · Porcine · Islet transplantation · Islets of Langerhans

Introduction

Although pancreas digestion using collagenase has be-come the “gold standard” for islet isolation, it is extremelyunpredictable [1]. One of the variables that could accountfor this is the collagen substrate [2] within the pancreas.Islet isolation has been shown to vary between species andit is generally agreed that porcine islet isolation is incon-sistent [3], whilst rodent islet isolation is somewhat moresuccessful [4]. It has previously been suggested [2, 5] thatthis, may be related to both quantitative and qualitativedifferences in the collagen content of porcine and rodentpancreata. The aim of the present study was to determinethe distribution of different collagen types within adult andjuvenile porcine pancreata.

S.A. White (✉) · D.P. Hughes · H.H. Contractor · N.J.M. LondonDepartment of Surgery, University of Leicester, PO Box 65, Clinical Sciences Building, Leicester, UK

J Mol Med (1999) 77:79–82 © Springer-Verlag 1999

ORIGINAL ARTICLE

S.A. White · D.P. Hughes · H.H. Contractor N.J.M. London

An investigation into the distribution of different collagen types within adult and juvenile porcine pancreata

Materials and methods

Animals. Five adult (2–3 years old) and 5 juvenile (less than 12 months) Large White porcine pancreata were assessed (DawkinsInternational, Congerstone, UK). The entire pancreas was resected atthe abattoir.

Pancreatic sampling. The pancreas was transported to the islet labo-ratory in cold hyperosmolar citrate (Travenol, Thetford, UK) andcleaned of surrounding connective tissue. Biopsies were taken from5 different areas; area 1=tail of splenic lobe, area 2=body of spleniclobe, area 3=head of splenic lobe, area 4=posterior lobe and area5=duodenal lobe.

Slide preparation and immunohistochemistry. Each biopsy was em-bedded and tissue sections (4 µm) cut. They were incubated for 45min in pepsin (Sigma, UK) (0.025% in HCL) at 37°C followed byincubation in 6% hydrogen peroxide for 10 min. Having washed thesection, normal rabbit serum (1:20) was overlaid for 10 min (Dako,Glostrup, Denmark). The primary collagen monoclonal antibodieswere then added. The collagen monoclonal antibodies were of col-lagen types I, V, VI (Southern Biotechnology Associates Inc, Bir-mingham, USA) and type IV (Dakopatts, Denmark). The sectionswere washed with PBS for 20 min and then overlaid with a peroxi-dase conjugated rabbit anti-goat antibody, or in the case of collagentype IV a biotinylated rabbit anti-mouse antibody for 30 min at roomtemperature (1:25) (Dako, Glostrup, Denmark). Those slides origi-nally overlaid with collagen type IV were then washed once againwith PBS and then incubated with a Vectastain Elite ABC (VectorLab, California, USA) immunoperoxidase staining protocol. Allslides were then dehydrated, cleaned and mounted. Other tissue sec-tions that were simultaneously prepared from adjacent tissue sec-tions were controls without the collagen monoclonal antibodies.These were stained alternatively with anti-insulin to allow identifica-tion of insulin containing islets of Langerhans.

Evaluation of collagen distribution. Each slide was randomized forassessment by two observers. The definition of collagen distributionwas assessed by a modification of that previously used by Van Suyli-chem et al. [2] (Table 1). Each slide was analysed by light microsco-py (Nikon, Diaphot, Japan) at ×40 magnification. Islets were identi-fied by tissue sections stained for insulin adjacent to thecorresponding collagen tissue section. The collagen content was as-sessed within the interlobar pancreatic septa, within the pancreaticlobules (intralobular), collagen surrounding the islet (peri-insular)and also the collagen within the islet (intra-islet capillary tissue).

Statistical analysis. Results are expressed as mean and 95% confi-dence intervals. Comparisons of all parameters between pairs were

performed using the Mann-Whitney U-test with its significance (ad-justed for ties) being expressed as a P value. Specific differences be-tween each of the four different collagens was initially analysed bythe Kruskall-Wallis one way analysis of variance test.

Results

Interlobar septa

Within the interlobar septa statistically significant differ-ences were apparent, between the different types of col-lagen in both adults (P<0.001) and juveniles (P<0.001).This was attributable to the differences between the in-tensely stained collagen type VI compared to the least in-tensely stained collagen type I in both adults (P<0.001)and juveniles (P<0.001). In contrast collagen types IV andV were moderately stained throughout but still significant-ly different to that of type VI (Fig. 1A).

Intralobular septa

This pattern was also consistent within the intralobular(acinar) septa where statistical differences were seen inboth adults (P<0.001) and juveniles (P<0.001). Collagentype VI was seen completely enveloping all acinar seg-ments with marked intensity, greater to that seen withinthe interlobar septa and more significant than collagentype I (P<0.001), type IV (P<0.001) and type V (P<0.001)in both adults and juveniles (Fig. 1B).

Peri-insular collagen

The peri-insular collagen content also showed diversitybetween collagen types but in the majority of both juvenileand adult pancreata, the peri-insular capsule was fragmen-tary and less than 50% of the total islet circumference.There were significant differences between collagen typesof both adult (P<0.001) and juvenile (P<0.001), this main-ly being attributable to the quantitative differences of col-

lagen types I (weak) and type VI (intense) in adults(P<0.001), and diversity between collagen type I, IV andV (moderate) compared to type VI (intense) in juveniles(P<0.001) (Fig. 1C). Moreover, the quantitative content ofcollagen within the islet peri-insular islet capsule showedan increasing trend in juvenile pancreata compared to thatseen adults.

Intra-islet

The intra-islet collagen content also showed significantdifferences in adults (P<0.001) and juveniles (P<0.001).Collagen type IV was as intense as type VI, with types Iand V being significantly less intense (Fig. 1D). All intra-islet collagen was identified within islet capillaries.

Discussion

One of the factors accounting for the fragility of porcineislets during the isolation procedure was thought to be theyoung age of porcine donors. Histologically, demarcationof islets from the surrounding acinar tissue is difficult be-cause of the absence of a fibrous capsule around the islet[6]. Subsequently larger islet yields can be obtained fromadult porcine pancreata [7]. The results in this study sug-gest that this discrepancy cannot be explained in terms ofcollagen content between adult and juvenile pancreata. Nodifference in either collagen type (I, IV, V, VI) could beobserved in the interlobar or intralobular septa, peri-insu-lar, or intra-islet regions in either juvenile or adult porcinepancreata, though their was a general trend towards an in-creased collagen content of juvenile peri-insular regions,particularly of collagen type VI. Nevertheless, the integrityof adult porcine islets in contrast to juvenile could be ex-plained by age where collagens from older individuals aremore resistant to degradation by bacterial collagenasethrough the formation of stable crosslinking [8].

These results support the findings of van Suylichem etal. [2], where no difference in the collagen content couldbe observed between juvenile and adult peri-insular re-gions. Furthermore, it was also suggested in the abovestudy, using sirus red absorbance, that the amount of col-lagen was significantly less within the pancreatic septa ofadult pancreas compared to juveniles indicating a loweramount of collagen specifically within interlobar and in-tralobular septa.

Collagen was identified within the peri-insular regionsurrounding up to 50% of the total area in most cases, con-tradictory to other studies [2, 5, 9, 10], though this over-sight could be attributable to the limitations of light micro-scopic techniques because of islet capillaries beingpositioned adjacent to acinar tissue apparently incorporat-ed within the peri-insular capsule. Peri-insular stainingwith collagen is known to be discontinuous and of inter-mediate intensity in other species such as dog and man, incontrast to that of pig which is of very low intensity andvery fragmentary being almost completely deficient of a

80

Table 1 Collagen scoring protocol

Collagen Distribution

Score Interlobar Intralobular Peri-insular Intra-islet(acinar)

Zero absent absent absent absent

One + + <50% circumference <50% total area

Two ++ ++ >50% circumference >50%–90% total area

Three +++ +++ totally surrounding >90% total area

peri-insular capsule with islets being in direct contact toacinar tissue through cell-cell adhesion [5].

The diversity of collagen content within the peri-insularcapsule of adult porcine islets has also been demonstratedby Ulrichs et al. [11] where at least 3 different collagentypes I, III and IV were identified. More recently Van De-ijnen et al. [12] described the distribution of collagen typesI, III,V and IV. The intensity of collagen type I in the peri-

insular region, interlobar and intralobular septa was weakin porcine species but moderate in other species such asrat, dog and man. In contrast collagen type III was mark-edly more intense though still weak in porcine species.That of type V also reacted weakly in pig, particularlywithin the interlobar and intralobular acinar pancreaticsepta but moderate in rat, dog and man. These findingswere also similar between the acinar cells but were muchless in the pig. The results in this study also complementthose findings, but indicate that collagen type VI predomi-nates in both adult and juvenile peri-insular regions [12].

The function of collagen type VI is not known but itsultrastructure suggests that it is an independent fibroussystem perhaps important to the development and mainte-nance of spatial separation of distinct tissue componentsfrom large banded collagen fibers [13]. The relevance ofthis collagen type still needs to be determined so as to al-low its exploitation by a commercial collagenase as it ap-

81

Fig. 1A–D Differences in collagen content of types I, IV, V, VI asdefined in text. Plots are mean and 95% confidence intervals. Fig-ures represent differences between all collagen types within juve-niles and adults using Kruskall Wallis one way analysis of variance.Comparison between pairs were then made, where valid, using theMann-Whitney U-test with its significance (adjusted for ties) beingexpressed as a P value. A Inter-lobar collagen distribution B Intra-lo-bar (acinar) collagen distribution. C Peri-insular collagen score DIntra-islet collagen score

A

B

C

D

pears to be the most abundant collagen type within porcinepancreata presented herein.

It can be seen that the assembly of collagens into acomplex fibrillar pattern within porcine pancreata is com-posed of many different types, at least types I, III, IV, Vand VI [11, 12]. This assortment obviously having con-straints upon the rate of degradation of native collagensubstrates through their different rates of individual degra-dation by collagenases [1]. It can therefore be assumedthat to minimize porcine islet fragmentation during theisolation procedure, a collagenase that does not degradeperi-insular collagen type VI would be ideal but the prob-lem is that collagen type VI is also abundant within theacinar tissue as well. In summary, it can be seen from thisstudy and others [2, 5, 9, 10] that the distribution of thedifferent collagen types and the content of the extracellularmatrix within pancreata has the potential to detrimentallyaffect islet isolation.

Acknowledgements We would like to thank the department of spe-cial histopathology, Leicester Royal Infirmary for helping to preparethe histological collagen sections.

References

1. Johnson PRV, White SA, London NJM (1995) Collagenase andhuman islet isolation. Cell Transplantation 3: 437–452

2. van Suylichem PTR, van Deijnen JHM, Wolters GHJ, vanSchilfgaarde R (1995) Amount and distribution of collagen inpancreatic tissue of different species in the perspective of isletisolation procedures. Cell Transplantation 4: 609–614

3. White SA, Contractor HH, Hughes DP, Johnson PRV, ClaytonHA, Bell PRF, London NJM (1996) The influence of differentcollagenase solvents and the timing of their delivery upon por-cine islet isolation. Br J Surg 83: 1356–1364

4. Lake SP, Anderson J, Chamberlain J, Gardner SJ, Bell PR,James RFL (1987) Bovine serum albumin density gradient isola-tion of rat pancreatic islets. Transplant 43: 805–808

5. van Deijnen JH, Hulstaert CE, Wolters GH, van Schilfgaarde R(1992) Significance of the peri-insular extracellular matrix forislet isolation from the pancreas of rat, dog, pig, and man. CellTissue Res 267: 139–146

6. Socci C, Ricordi C, Davalli AM, Staudacher C, Baro P, VertovaA, Freschi M, Gavazzi F, Braghi S, Pozza G, Di Carlo V (1996)Selection of donors significantly improves pig islet isolation.Horm Met Res Suppl 25: 32–34

7. Ricordi C, Socci C, Davalli AM, Staudacher C, Baro P, VertovaA, Sassi I, Gavazzi F, Pozza G, Di Carlo V (1990) Isolation ofthe elusive pig islet. Surgery 107: 688–694

8. Weiss JB (1982) Degradation of collagen. In: Weiss JB, JaysonMV (eds) Collagen Health and Disease. Churchill Livingstone,Edinburgh p 126

9. Uscanga L, Kennedy RH, Stoker S, Grimaud JA, Sarles H(1984) Immunolocalisation of collagen types, laminin andfibronectin in normal human pancreas. Digestion 30: 158–164

10. Wolters GHJ, Vos-Scheperkeuter GH, van Deijnen JHM, vanSchilfgaarde R. An analysis of the role of collagenase and pro-tease in the enzymatic dissociation of the rat pancreas for isletisolation (1992). Diabetologia 35: 735–742

11. Ulrichs K, Bosse H, Wacker H, Heiser A, Müller Ruchholtz W(1994) Histologic analysis of the porcine pancreas to improve is-let yield and integrity after collagenase digestion. TransplantProc 26: 610–612

12. van Deijnen JHM, van Suylichem PTR, Wolters GHJ, van Schilf-gaarde R (1994) Distribution of collagens type I, type III, TypeIV and type V in the pancreas of rat, dog, pig and man. Cell Tis-sue Res 277: 115–121

13. Burgeson RE, Nimni ME (1992) Collagen types. Molecularstructure and tissue distribution. Clin Ortho Rel Res 282:250–272

82