Correspondence
Identification of the Whipple's Disease Bacillus
N Engl J Med 1993; 328:62-63January 7, 1993
- Article
To the Editor:
We were puzzled by the recent report by Relman et al. (July 30 issue)1 confirming the identification of the Whipple's disease bacillus. A year ago in The Lancet we also reported that the Whipple bacillus was a gram-positive bacterium rich in guanine and cytosine (an actinomycete) not closely related to any other known organism2. Our report showed that the biopsy specimen of small intestine from a patient with Whipple's disease contained massive numbers of a bacterial agent with the same 16S rRNA sequence as that confirmed and extended by Relman et al. We later reported using a diagnostic polymerase chain reaction (PCR) to show the same sequence in the lymph node of a patient with only extraintestinal Whipple's disease3. The paper by Relman et al., and even more, the accompanying editorial,4 gave the impression that our findings were somehow only preliminary. Our group certainly would not describe as preliminary the report in the Journal by Relman et al. describing the phylogeny of the agent of bacillary angiomatosis on the basis of the analysis of only 480 bases5. Yet the study of Rochalimaea henselae has continued to progress6. Science always builds on previous work.
The authors have clearly overstressed the differences between our previous work and their own. Nowhere did we suggest, as they stated, that the organism was a nocardioform. As compared with other rRNA sequences available at that time, the Whipple bacillus had slightly more sequence similarity to Rhodococcus equi (a nocardioform) than to Arthrobacter globiformis (an actinobacterium) or Streptomyces lividans (a streptomycete). However, R. equi is a slowly evolving organism that remains similar to all actinomycetes and their common ancestor.
Casual readers may conclude that the authors' phylogenetic tree settles the evolutionary history of the Whipple bacillus. To their credit, they make no such claim. Bootstrap analysis, a measure of statistical confidence, associated the Whipple bacillus with the actinobacteria only 67 percent of the time -- far from the 95 percent confidence level required for a scientific conclusion. Thus, the Whipple bacillus may still represent, as we suggested, a fourth line of descent within the actinomycetes. Its relation to R. equi remains relevant because the latter organism uniquely causes a histologic appearance similar to Whipple's disease. This unique phenotypic property shared with a “primitive” organism such as R. equi is also compatible with the theory that the Whipple bacillus represents a separate lineage. Where is the reservoir? What similar organisms are out there? If examples of such organisms were found, placement of the Whipple bacillus might become more definite.
6 ReferencesKenneth H. Wilson, M.D.
Rhonda B. Blitchington, B.S.
Richard Frothingham, M.D.
Veterans Affairs Medical Center, Durham, NC 27705Joanne A.P. Wilson, M.D.
Duke University Medical Center, Durham, NC 277101
Relman DA ,Schmidt TM ,MacDermott RP ,Falkow S . Identification of the uncultured bacillus of Whipple's disease. N Engl J Med 1992;327:293-301
Full Text | Web of Science | Medline2
Wilson KH ,Blitchington R ,Frothingham R ,Wilson JAP . Phylogeny of the Whipple's-disease-associated bacterium. Lancet 1991;338:474-475
CrossRef | Web of Science | Medline3
Wilson KH . New vistas for bacteriologists: analysis based on 16S rRNA sequences provides a rapid and reliable approach to the identity of human pathogens. ASM News 1992;58:318-3214
Donaldson RM Jr . Whipple's disease -- rare malady with uncommon potential. N Engl J Med 1992;327:346-348
Full Text | Web of Science | Medline5
Relman DA ,Loutit JS ,Schmidt TM ,Falkow S ,Tompkins LS . The agent of bacillary angiomatosis -- an approach to the identification of uncultured pathogens. N Engl J Med 1990;323:1573-1580
Full Text | Web of Science | Medline6
Regnery RL ,Anderson BE ,Clarridge JE III ,Rodriguez-Barradas MC ,Jones PC ,Carr JH . Characterization of a novel Rochalimaea species, R. henselae sp. nov., isolated from blood of a febrile, human immunodeficiency virus-positive patient. J Clin Microbiol 1992;30:265-274
Web of Science | Medline
To the Editor:
In their generally excellent report, Relman et al.1 fail to demonstrate that the proposed causal organism is in fact localized to the pathologic lesion. This could have been shown, simply and definitively, by in situ hybridization with a probe specific for that organism. It is surprising, given the care the authors otherwise showed, that they did not include such an obvious experiment, since it would have greatly strengthened their conclusion -- which, though compelling, does not rest on fulfilling Koch's postulates.
Given the emphasis placed on the generality and importance of the approach used, it seems incumbent on us to point out that, although it is unacknowledged in the article, we in fact conceived and developed this approach2. Specifically, we proposed the use of the PCR with broad-range eubacterial 16S ribosomal primers, not cross-reactive with human sequences, to recognize and classify uncultured bacterial pathogens in idiopathic human disease, precisely as was done here and in a similar previous report3. In this work we extensively characterized broad-range oligonucleotide primers suitable for this purpose and used by Relman et al. in slightly modified form. The approach was based in part on our earlier use, probably for the first time, of bacterial nucleic acid probes to detect bacteria in eukaryotic tissue for the study of uncultured bacteria in disease4. Thus, the authors' statement,1 “We have developed a technique for . . .” is somewhat misleading. More accurately, these reports represent fine applications of this technique that, gratifyingly, validate our original expectations.
In developing this approach, our goal has been to concentrate on such important diseases as the idiopathic granulomatoses and the chronic arthritides, in which only small numbers of organisms, if any, would be expected. This contrasts markedly with the disorders studied by Relman et al., in which the organisms are so numerous as to be readily visualized microscopically. Thus, in our studies the extraordinary sensitivity of the PCR can be better exploited, although the accompanying problem of contamination requires a much more stringent solution. For diseases in which a bacterial cause remains uncertain, this approach has the important potential, in the event of a negative result (assuming adequate sensitivity and controls), of providing substantial evidence against the presence of bacterial DNA and thus against a bacterial cause. Such evidence cannot be provided by other known methods.
4 ReferencesCharles R. Steinman, M.D.
Peter Rumore, M.D.
State University of New York at Stony Brook, Stony Brook, NY 117941
Relman DA ,Schmidt TM ,MacDermott RP ,Falkow S . Identification of the uncultured bacillus of Whipple's disease. N Engl J Med 1992;327:293-301
Full Text | Web of Science | Medline2
Chen K ,Neimark H ,Rumore P ,Steinman CR . Broad range DNA probes for detecting and amplifying eubacterial nucleic acids. FEMS Microbiol Lett 1989;48:19-24
CrossRef | Web of Science | Medline3
Relman DA ,Loutit JS ,Schmidt TM ,Falkow S ,Tompkins LS . The agent of bacillary angiomatosis -- an approach to the identification of uncultured pathogens. N Engl J Med 1990;323:1573-1580
Full Text | Web of Science | Medline4
Steinman CR . Specific detection and semiquantitation of micro-organisms in tissue by nucleic acidihybridization. I. Characterization of the method and application to model systems. J Lab Clin Med 1975;86:164-174
Medline
Author/Editor Response
The authors reply:
To the Editor: Steinman and Rumore point out the usefulness of in situ hybridization to link physically the bacterial 16S rRNA sequence we amplified from Whipple's disease tissues with visible bacterial forms. We agree that in situ hybridization might help to strengthen our findings. Nonetheless, we believe that the use of a substantial number of properly chosen control tissues may allow one to make a meaningful, specific association of a bacterial 16S rRNA sequence with a clinicopathological syndrome1. Despite the obvious relevance of in situ hybridization to this type of investigation, the use of oligodeoxynucleotide probes with formalin-fixed tissue can present technical difficulties. We are currently working to solve these problems for both in situ hybridization and in situ PCR.
Concerning the origins of this experimental approach, we acknowledge the proposals of Chen and coworkers2 and have cited them in an earlier paper3. However, the concept of using broad-range rRNA PCR primers to identify uncultured organisms had been suggested4 even before the work of Chen et al. We believe that our work illustrates the assimilation of these suggestions and various techniques and their application to practice.
We clearly acknowledged the work of Wilson and coworkers5 in our report and stated that their partial (<50 percent) 16S rRNA sequence was identical to a portion of our reported sequence, However, we wish to reiterate our belief that the work of Wilson et al. was preliminary. Their investigation involved only one patient with Whipple's disease and no control tissues. They did not show a specific association between this sequence and tissues with Whipple's disease. Their claim of demonstrating the Whipple bacillus sequence in a second patient was not presented in a peer-reviewed journal, nor were any details of this work provided. We delayed the presentation of our data for over one year, despite obtaining several positive results in patients with Whipple's disease and more than 50 percent of the 16S rRNA gene sequence. We waited until we had analyzed 14 control tissues from 10 patients. The importance of controls in this kind of work cannot be overemphasized6. In addition, we proposed and tested reagents for the specific detection of this organism. As for the use of a partial sequence, in our initial investigation of the agent of bacillary angiomatosis3 we first established that the 480 base positions of 16S rRNA sequence were sufficient for recreating the phylogenetic relations among the α-proteobacteria. We only published our findings after multiple tissues from patients with the disease and from controls were analyzed. The analysis of the entire 16S rRNA gene sequence later confirmed these results7.
Our sequence data are more complete and therefore make possible a more definitive analysis of the ancestry of the Whipple bacillus. We acknowledge that Wilson et al. did not state that the Whipple bacillus was a nocardioform. However, some readers may have come to this conclusion, since R. equi (a nocardioform) was given the highest score for sequence similarity. Finally, we agree that the final word has not been written on the evolutionary relation of the Whipple bacillus (Tropheryma whippelii). With the availability of more 16S rRNA sequences from related organisms, our appreciation of the evolutionary descent of T. whippelii will continue to improve.
7 ReferencesDavid A. Relman, M.D.
Stanford University, Stanford, CA 94305Richard P. MacDermott, M.D.
University of Pennsylvania, Philadelphia, PA 19104Thomas M. Schmidt, Ph.D.
Miami University, Oxford, OH 450561
Relman DA, Falkow S. Identification of uncultured microorganisms: expanding the spectrum of characterized microbial pathogens. Infect Agents Dis (in press).
2
Chen K ,Neimark H ,Rumore P ,Steinman CR . Broad range DNA probes for detecting and amplifying eubacterial nucleic acids. FEMS Microbiol Lett 1989;48:19-24
CrossRef | Web of Science | Medline3
Relman DA ,Loutit JS ,Schmidt TM ,Falkow S ,Tompkins LS . The agent of bacillary angiomatosis -- an approach to the identification of uncultured pathogens. N Engl J Med 1990;323:1573-1580
Full Text | Web of Science | Medline4
Medlin L ,Elwood HJ ,Stickel S ,Sogin ML . The characterization of enzymatically amplified eukaryotic 16S-like rRNA-coding regions. Gene 1988;71:491-499
CrossRef | Web of Science | Medline5
Wilson KH ,Blitchington R ,Frothingham R ,Wilson JAP . Phylogeny of the Whipple's-disease-associated bacterium. Lancet 1991;338:474-475
CrossRef | Web of Science | Medline6
Donaldson RM Jr . Whipple's disease -- rare malady with uncommon potential. N Engl J Med 1992;327:346-348
Full Text | Web of Science | Medline7
Relman DA ,Lepp PW ,Sadler KN ,Schmidt TM . Phylogenetic relationships among the agent of bacillary angiomatosis, Bartonella bacilliformis, and other α-proteobacteria. Mol Microbiol 1992;6:1801-1807
CrossRef | Web of Science | Medline
- Citing Articles (3)
Citing Articles
1
DENIS DURAND, CLAIRE LECOMTE, PASCAL CATHEBRAS, HUGUES ROUSSET, PIERRE GODEAU. (1997) Medicine 76:3, 170
CrossRef2
Lowsky, RobertArcher, Gordon L.Fyles, GillianMinden, MarkCurtis, JohnMessner, HansAtkins, HaroldPatterson, BruceWilley, Barbara M.McGeer, Allison. (1994) Diagnosis of Whipple's Disease by Molecular Analysis of Peripheral Blood. New England Journal of Medicine 331:20, 1343-1346
Full Text3
Bethi Muralidhar, Peter M. Rumore, Charles R. Steinman. (1994) Use of the polymerase chain reaction to study arthritis due toneisseria gonorrhoeae. Arthritis & Rheumatism 37:5, 710-717
CrossRef
