Recently, the G191R variant in anionic trypsinogen (coded by PRSS2) was reported to be strongly associated with chronic pancreatitis (CP) (odds ratio (OR) = 0.37, p = 1.1×10⁻⁸) in European patients with CP1 and subsequently replicated in Hungarian and Japanese patients with CP.2 3 We and others have earlier shown that CP in tropical countries such as India and Bangladesh, especially tropical calcific pancreatitis (TCP), has a variable genetic basis compared to that in Europeans and rather than cationic trypsinogen (coded by PRSS1) mutations,3 4 variants in SPINK1 encoding the pancreatic secretory trypsin inhibitor or cathepsin B (CTSB) or chymotrypsin C predict susceptibility to TCP.3^–7 We, therefore, investigated using a case–control approach whether G191R PRSS2 variant also confers protection against TCP.

We analysed the G191R variant in the PRSS2 gene by direct sequencing in 700 patients with CP including 500 patients with TCP, 155 patients having idiopathic chronic pancreatitis (ICP) and 45 with alcoholic pancreatitis (see supplementary material). On initial screening of 299 patients with TCP and 277 control subjects, the minor allele frequency (MAF) of the G191R variant was significantly lower compared to Europeans1 and was similar in both the groups (0.3%). Genotyping and pooled analysis of additional patients with CP and normal subjects showed comparable frequency of the protective allele 191R in patients (MAF = 0.7% (5/700)) vs 0.3% in normal subjects (4/1272) and suggested no significant association with CP ((OR = 2.27; 95% confidence interval, 0.53 to 10.05; p = 0.21) (supplementary table 1a)). All the individuals carried the G191R variant in heterozygous state. Further genotyping of five SNPs (rs10273639, rs1985888, rs2855983, rs997222 and rs2367487) around PRSS2 showed comparable D′ values (approximately 0.90) and the same haplotype background as observed in the Europeans, suggesting a recent origin for the R191 allele.1 (supplementary table 1b). Interestingly, the MAF at G191R in our study is comparable to that observed in Africans.1 It is believed that migrations out of Africa, after passing through a population bottleneck, led to African–Eurasian divergence around 100 000 years ago, followed by an European Asian divergence about 40 000 years ago (http://www.actionbioscience.org/evolution/johanson.html). So far, no PRSS1 or PRSS2 variants, including the recently reported copy number mutation at the trypsinogen locus, have been identified in TCP or ICP patients from India and other tropical countries.4^–9 Enough evidence exists to suggest that CP in Asian Indians may be caused through a different mechanism,4^–9 a fact supported by lack of association of PRSS1 variants with TCP in Indians and CTSB variants with ICP in Europeans.10 In addition, the content of digestive enzymes such as trypsin and chymotrypsin in the pancreatic juice of Indians is two to four times lower than in Europeans,11 probably, as a consequence of a lower content of the delimiting amino acids such as methionine, lysine and tryptophan in their predominantly vegetarian diet, implying a limited role for anionic trypsinogen in pancreatic protein digestion.12 Thus, it may be surmised that the protective PRSS2 variant has only marginally penetrated the Indians possibly because it does not play an important role in protection against TCP. Although a recent study speculated a protective role to G191R polymorphism in TCP based on analysis of 174 patients and 794 control subjects from north and south India, the study did not reach statistical significance (p = 0.48).6 Considering the small sample size of north Indian control subjects (n = 84), any assumptions on role of differing allele frequency for G191R variant in various parts of India remains a matter of debate.6

The above results assume immense importance since by in vitro analysis of activation of the recombinant G191R protein, Witt et al have suggested that it protects against CP by mitigating intrapancreatic trypsin activity.1 We can only speculate about the possible explanations for our findings. Although mechanisms similar to cationic trypsinogen are extended to anionic trypsinogen in the pathogenesis of CP, no gain-of-function variant in PRSS2 has been identified so far. Of the three isoforms of trypsinogens secreted by the human pancreas, the cationic isoform constitutes about 70–75% and anionic trypsinogen makes up only about 25–30%. Hence, the larger pool of cationic trypsinogen would remain available for autoactivation, even after the autocatalytic degradation of G191R anionic trypsinogen. Additionally, Kukor et al13 have shown that autoactivation of anionic trypsinogen is inhibited at the acidic pH in the intracellular vesicular compartments that are known sites of pathological trypsinogen activation, while Ca²⁺ levels influence the autoactivation and stability of both cationic and anionic trypsinogens. Taken together, cationic trypsinogen remains the major isoform that determines the rate of activation and levels of trypsin generated and anionic trypsinogen may be of little consequence in the initiation of TCP through premature activation of trypsinogens. Thus, inactivation of G191R anionic trypsinogen may have minor protective effect, which was not seen in TCP, possibly affected by other genetic and environmental factors.

One of the potential limitations of our study could be the limited power (27.5%) to detect the association (http://www.dssresearch.com/toolkit/spcalc/power_P2.asp). However, to achieve at least 80% power with the present MAF for G191R variant necessitates a sample size of about 5000 cases and controls each, which is fairly difficult to achieve for relatively rare diseases such as TCP. Our study was 82% powered to detect an association assuming an OR and allele frequency as observed in the original study (OR = 0.37; MAF = 1.3% in patients and 3.4% in controls).1

To conclude, we did not find any association of the G191R PRSS2 variant in Indian patients with CP, especially those with TCP, thus adding to our earlier observation that trypsinogen variants may not play an important role in the pathogenesis of CP in Indians.

Acknowledgements

The authors thank all the subjects who gave informed consent and voluntarily participated in the study. The authors are also grateful to Drs T Ramakrishna Murti and Yogendra Sharma, CCMB, Hyderabad; and Ms S Esther, Asian Institute of Gastroenterology, Hyderabad for helpful discussions on the manuscript.