Brimming with b's. Newfound cells in the pancreas give rise to neurons (red) and insulin-producing b cells (green).
每年的11月14日是世界糖尿病日，糖尿病在全世界发病率相当高，现在全球糖尿病病人已超过1.7亿，且每年都在快速增长。糖尿病是继心血管和肿瘤之后的第三大“健康杀手”。
我们的血液中含有一定浓度的糖（血糖），为人体提供能量。若血糖浓度过高或过低，就会引起疾病。而保持血糖浓度的功臣就是胰岛素，它来源于胰腺，由胰岛内的一种叫β细胞产生，并释放入血液。
近几年，1型糖尿病患者在逐年增加，越来越年轻化。1型糖尿病是一种自身免疫性疾病，由于体内多种T细胞（CD4＋细胞毒T淋巴细胞和CD8＋细胞毒T淋巴细胞）对胰岛β细胞抗原发生反应，引发白细胞浸润，从而产生炎性细胞因子和自由基，导致胰岛β细胞死亡。胰岛素水平降低，导致血糖浓度升高。
因此，对如何增加胰岛β细胞数量的研究就成了研究者们致力的难题。刚开始，研究者们就考虑到了利用胰腺干细胞可能会分化为β细胞，但研究发现β细胞是由其他类型的β细胞产生的，因此就停止了对胰腺干细胞的进一步研究。
今年8月24日Nature说，胰腺干细胞很有可能产生大量的β细胞。一个研究小组在成年小鼠胰腺内鉴定了一组细胞群可以分化产生大量的β细胞。多伦多大学的Derek试验小组用促进神经干细胞生长的传统培养基浸洗这一群细胞，发现每5000个细胞中就由一个细胞能够迅速地繁殖成一组细胞群，且这群细胞具有不同于典型细胞的特性。
这群细胞还有一个重要特点:它们能够很自然地发育成不同的组织。当改变培养基且促进细胞分化时，可以产生包括β细胞在内的大量的多种类型的胰腺细胞。若在培养基中加入糖，β细胞能够释放出高于平常两倍的荷尔蒙，这就是体内β细胞通过释放胰岛素对糖作出反应。
Derek试验小组在8月22日的Nature生物技术上发表了他们的研究内容。但他们还没有证明这一群细胞就是胰腺干细胞，不过这已经在糖尿病长期治疗的路上又迈了一步。不管怎样，这都是一个很大的突破，具有重大的意义。
Clonal identification of multipotent precursors from adult mouse pancreas that generate neural and pancreatic lineages
The clonal isolation of putative adult pancreatic precursors has been an elusive goal of researchers seeking to develop cell replacement strategies for diabetes. We report the clonal identification of multipotent precursor cells from the adult mouse pancreas. The application of a serum-free, colony-forming assay to pancreatic cells enabled the identification of precursors from pancreatic islet and ductal populations. These cells proliferate in vitro to form clonal colonies that coexpress neural and pancreatic precursor markers. Upon differentiation, individual clonal colonies produce distinct populations of neurons and glial cells, pancreatic endocrine -, - and -cells, and pancreatic exocrine and stellate cells. Moreover, the newly generated -like cells demonstrate glucose-dependent Ca2+ responsiveness and insulin release. Pancreas colonies do not express markers of embryonic stem cells, nor genes suggestive of mesodermal or neural crest origins. These cells represent a previously unidentified adult intrinsic pancreatic precursor population and are a promising candidate for cell-based therapeutic strategies.
Figure 1. PMP colonies are formed from progenitors present in adult pancreatic islet and duct cell isolates, and express markers characteristic of both neural and pancreatic precursors.
(a) The frequency of PMP colonies from pancreatic islet and duct cell isolates is similar. The data are expressed as the mean number of colonies (+s.e.m.; n = 14 independent experiments) formed per 10,000 cells plated. Islet and ductal cell isolates do not contain significantly different numbers of PMPs (P > 0.05). (b) Light micrograph of a PMP colony. Scale bar, 50 m. (c) Light micrograph of a neurosphere. Scale bar, 50 m. (d) RT-PCR for neural and pancreatic precursor markers. The numbers on the left represent the number of individual PMP colonies that expressed the corresponding mRNA out of the total number of colonies tested by RT-PCR analysis. Only single colony RNA isolates that were found to express -actin were considered. Note that positive control (+) bands (see Supplementary Methods online for a complete list of tissue positive controls) appear brighter because of the greater amount of starting RNA in comparison to single PMP colonies. Ngn3 was not expressed at detectable levels in individual PMP colonies. However, Ngn3 mRNA was detected in a sample of five pooled (P) PMP colonies, suggesting that it is present in differentiated PMP colonies but perhaps at low levels. (e) Single cells from dissociated PMP colonies coexpress PDX-1 (red) and nestin (green) as seen by immunostaining. Note that the nucleus in this fluorescence micrograph is labeled with both DAPI (blue) and PDX-1, giving it a pink appearance. The white arrows indicate double-positive cells.
Figure 2. PMP colonies generate all three major neural cell lineages.
(a,b) When individual PMP colonies were differentiated, they were found to generate 3-tubulin+ neurons (red), occasionally forming large neuronal networks as shown in b. Scale bars: 50 m, a; 200 m, b. (c–e) 3-tubulin+ neurons that were generated by PMPs (c) coexpressed the more mature neuronal marker MAP2 (green) (d, and overlay e), thus confirming their neuronal identity. Scale bar, 50 m. (f,g) PMPs generated GFAP+ astrocytes (green). Scale bars, 20 m. (h) O4+ oligodendrocytes also were generated by PMP colonies (green). Scale bar, 20 m. All nuclei were counterstained with DAPI (blue) for purposes of quantification. Refer to Table 1 for relative proportions of each neural cell type produced by PMPs. (i) RT-PCR analyses confirm the presence of mRNA for neuronal and glial makers. Individual differentiated clonal PMP colonies all expressed detectable levels of 3-tubulin and MAP2, but not GFAP. However, GFAP mRNA was detected in a sample of five pooled (P) PMP colonies, suggesting that it is present in differentiated PMP colonies but at lower levels. This is in accordance with the relatively lower percentages of glial than neuronal progeny determined by immunocytochemistry (Table 1). Only single colony RNA isolates that were found to express -actin were considered. Note that positive control (+) bands appear brighter because of the greater amount of starting RNA in comparison to single PMP colonies.
Figure 3. Progeny from two distinct embryonic primary germ layers are generated by single, clonally derived PMPs that are present in islet and ductal cell isolates.
(a,b) Upon differentiation, single islet- (a) and ductal- (b) derived PMP colonies generated both 3-tubulin+ neurons (red) and insulin+ or C-peptide+ -cells (green). Note that although only one combination of 3-tubulin and insulin or C-peptide is shown for each of islet and ductal PMP colonies, both islet and ductal PMP colonies contained insulin+ and C-peptide+ cells in combination with 3-tubulin. The white arrows indicate insulin+ and C-peptide+ cells. Scale bars, 50 m. (c,d) To confirm that the insulin+ cells represented -cells and were generating insulin protein de novo, differentiated colonies were colabeled with antibodies against PDX-1 and C-peptide (c) or insulin (d). These micrographs illustrate single colonies with cells positive for both PDX-1 (red) and C-peptide or insulin (green). Scale bars, 25 m. (e,f) Insulin+ cells (red) all coexpress C-peptide (green) as illustrated by the merged field (yellow) (e) and C-peptide+ cells (green) all coexpress Glut2 (red) as shown in the merged field (yellow) (f). Scale bars, 50 m. Although only one example of each is illustrated, both islet- and ductal-derived PMP colony progeny exhibited these patterns. In all micrographs nuclei have been counterstained with DAPI for purposes of quantification. Note that in c and d, nuclei appear pink because of the colocalization of DAPI and PDX-1. Refer to Table 1 for the proportion of cells with -cell characteristics produced by single PMPs. (g) RT-PCR analyses confirm that single clonal differentiated PMP colonies express many characteristic islet/-cell markers. Only single-colony RNA isolates that were found to express -actin were considered. Note that positive control (+) bands appear brighter because of the greater amount of starting RNA in comparison to single PMP colonies.
Figure 4. Insulin+ cells generated de novo from PMPs demonstrate glucose-stimulated Ca2+ responses and glucose-stimulated insulin release.
(a,b) Bright-field and fluorescence micrographs demonstrating YFP+ cells from AdRIP2EYFP-infected islet- (a) and ductal- (b) derived differentiated PMP colonies. Scale bars, 50 m. (c,d) Calcium traces for islet- (c) and ductal- (d) derived PMP colonies demonstrating glucose-stimulated [Ca2+]i responses, which were augmented by the addition of either GLP-1 or TEA, respectively. The addition of the voltage-dependent Ca2+ channel blocker verapamil (VER) returned the [Ca2+]i to basal levels. Shown above the Ca2+ trace are fluorescence micrographs of YFP+ cells and the ratiometric Fura images (pseudocolored according to the scale shown to the right) corresponding to the numbered time points on the trace. Note that in (c), the YFP- cell does not demonstrate a glucose response. These Ca2+ traces are representative of at least five independent experiments. Note that GLP-1 and TEA produced similar responses in both islet- and ductal-derived PMP progeny, although only one example is depicted for each in the [Ca2+]i traces shown. (e,f) Demonstration of increased insulin release by islet- (e) and ductal- (f) derived PMP colonies in response to high glucose (HG) alone or with the addition of GLP-1, TEA or to Carbachol (Carb) alone. The addition of verapamil (VER) abolished the glucose-stimulated insulin release. These data were generated from three to four independent experiments.
Figure 5. PMP colonies generate multiple islet endocrine subtypes and exocrine cells.
(a) When individual PMP colonies were differentiated, they were found to generate glucagon+ -cells (green) and somatostatin+ -cells (red). Cells coexpressing these hormones were never observed. Note that this field depicts only a portion of a larger differentiated PMP colony. The arrangement of endocrine cells in these colonies is suggestive of either multiple divisions of one local progenitor cell within the colony, or that there may be a type of 'community effect' whereby endocrine cells of similar phenotype tend to differentiate in close contact with each other. (b) PMP colonies generated cells characteristic of the exocrine compartment of the pancreas, amylase+ acinar cells. (c-d) A large proportion of the cells generated by individual clonal PMP colonies were large, flat cells with characteristic morphology and arrangement that expressed SMA (c) and nestin (d), typical of pancreatic stellate cells. All nuclei were counterstained with DAPI (blue) for purposes of quantification. Refer to Table 1 for relative proportions of each pancreatic cell type produced by PMPs. Scale bars, 25 m.
Figure 6. PMPs are not general endodermal or mesodermal precursors, nor are they ES cell–like stem cells or neural crest precursors.
(a) Individual PMP colonies were assayed by RT-PCR for the presence of the early endoderm markers GATA-4 and HNF3. None of the colonies tested expressed either marker, suggesting that PMPs are not generalized endodermal precursors. (b) mRNA for Oct4 and Nanog, proteins encoded by genes characteristic of ES cells, was not detected in any of the single clonal PMP colonies assayed, suggesting that PMPs are not ES cell–like pluripotent stem cells. (c) Brachyury and GATA-1, markers of mesodermal tissue, were not detected by RT-PCR in PMP colonies, suggesting that PMPs are not of mesodermal origin. (d) Clonal PMP colonies do not exhibit a characteristic neural crest progenitor profile. Although PMP colonies do express Slug and Snail, and a proportion of them express detectable levels of p75, they do not express many other characteristic neural crest markers including Pax3, Twist, Sox10 or Wnt1 by RT-PCR analysis. Only single colony RNA isolates that were found to express -actin were considered. Note that positive control (+) bands appear brighter because of the greater amount of starting RNA in comparison to single PMP colonies.
