HISTORY

In 1907, the pathologist George H. Whipple reported in detail, the case of a 36-year-old male physician-missionary who died after a five-year illness involving arthritis, chronic cough, weight loss, and chronic diarrhea.¹ At autopsy, Whipple found lipid deposits in the intestinal mucosa as well as in mesenteric and retroperitoneal lymph nodes. Microscopic examination further revealed a large number of macrophages with foamy cytoplasm in the lamina propria of the small intestine. Whipple suspected a disorder of fat metabolism and proposed the term intestinal lipodystrophy for the disease subsequently bearing his name.

In the following decades, only a few cases were reported and the diagnosis uniformly was made at autopsy. The first antemortem diagnosis was made in 1947 based on findings in a mesenteric lymph node removed at laparotomy,² and the first diagnosis by peroral intestinal biopsy was made in 1958.³ In 1949, Black-Schaffer⁴ introduced the periodic acid-Schiff (PAS) stain to the histopathologic diagnosis of Whipple’s disease. Inclusions in macrophages stained red using this stain, thus documenting that the intracellular material was glycoprotein rather than lipid.

The first report of successful antibiotic treatment (using chloramphenicol) was published in 1952.⁵ In 1961, two groups independently visualized bacteria by electron microscopy in affected tissues^6,7; subsequent reports confirmed these observations. The bacteria associated with Whipple’s disease are rod-shaped and of uniform size. Consistent positive therapeutic effects were achieved with antibiotic treatment.⁸ These findings and the positive PAS reaction⁴ suggested that the disease was unlikely to be a primary disorder of fat metabolism, but rather that it was a bacterial disease; efforts to cultivate this bacterium before 2000 failed to yield reproducible or consistent results.

The nature of the bacterium remained obscure until the early 1990s, when its 16S ribosomal DNA (rDNA) sequence was determined and phylogenetic analysis established the relationship of the bacterium to the actinomycetes.^9,10 The name Tropheryma whippelii was introduced,¹⁰ and the novel 16S rDNA sequence provided the basis for sensitive diagnostic testing using the polymerase chain reaction (PCR). In situ hybridization experiments showed that the unique bacterial 16S rRNA sequence colocalized with areas of pathology, thus supporting the relevance of the sequence and organism, the presence of which was thereby inferred.¹¹ Further advancement came with successful propagation of the Whipple’s disease bacterium in coculture with human fibroblast cells.¹² At that point, the bacterium formally was described as a new species, and its name was modified to Tropheryma whipplei.¹³ With the availability of adequate amounts of purified genomic DNA, the complete genome sequences of two different bacterial isolates were determined and published in 2003.^14,15

EPIDEMIOLOGY

Whipple’s disease is a rare disorder. The first comprehensive epidemiologic survey was performed by Dobbins in 1987,⁸ compiling information on 696 patients and including 617 published and 79 unpublished cases recorded through 1986. According to this analysis, Whipple’s disease is a sporadic disorder with a predilection for middle-aged white men. Data on age and sex were available for 664 patients; 86% were male, and the mean age at diagnosis was 49 years. Most patients were white; only 10 were African, one was a Native American, three patients were from India, and one was Japanese. Most of the patients originated from Europe (373 patients) or from the United States (246 patients). Within Europe, Germany (114 patients) and France (91 patients) were strongly represented. Relatively few cases originated from South America (11 patients) and Australia (13 patients).

A small epidemiologic study from western Switzerland calculated the incidence of Whipple’s disease to be approximately 0.4 per million of the population per year.¹⁶ A similar incidence of 0.4 per million per year was estimated for Germany.¹⁷ An epidemiologic analysis of 110 patients with Whipple’s disease in Germany, identified between 1965 and 1995, noted the incidence of cases to be relatively stable over three decades and a relatively even geographic distribution of the patients’ residences.¹⁸ There are only a few observations of geographically confined case clusters (up to seven cases).^19–21

In recent decades, several studies have indicated a statistically significant increase in the age of patients at diagnosis and an increasing percentage of female patients.^18,22,23 Currently, patients are first dianosed at a mean age of 56 years.¹⁸ It has been speculated that the increasing use of antibiotics for unrelated complaints may be a contributing factor in delaying the age of onset of Whipple’s disease. In a cohort of 191 patients whose Whipple’s disease was diagnosed between 1992 and 2007, 75% were male and 25% were female.²⁴ There are virtually no cases in children and young adults.

One remarkable epidemiologic feature in Dobbins’ analysis⁸ was the strong representation of patients with occupations in the farming and building trades, involving outdoor work or frequent contact with animals or soil: of 191 patients for whom data were available, 43 (22%) were farmers and 10 (5%) were carpenters; patients in all farming-related trades accounted for 34% of the total; by comparison, the fraction of farm workers among the total workforce in the analyzed countries was approximately 10%.

MICROBIOLOGY AND GENOMICS

After many unsuccessful attempts to cultivate the bacterium associated with Whipple’s disease, successful propagation of T. whipplei was reported in 2000, using infected heart valve tissue in coculture with human fibroblast cells.¹² Seven culture passages were performed, over a period of 285 days, and the 16S rDNA sequence of T. whipplei was detected after each passage. The initial estimate of the doubling time of the bacterium was 18 days, which represents extremely slow growth. Since the initial report, approximately 20 to 30 additional strains of T. whipplei have been isolated from various clinical specimens, including infected heart valves, duodenal biopsy specimens, ocular vitreous fluid, cerebrospinal fluid (CSF), synovial fluid, heparinized blood, mesenteric lymph node tissue, muscle tissue, and feces.^25–31 Methods to determine growth and identity of the bacterium in culture include immunofluorescence,^25,28 nucleic acid staining (Fig. 106-1),²⁹ endpoint PCR and sequencing,^25,28,29 quantitative PCR,^27,29 electron microscopy (Fig. 106-2),²⁹ and in situ hybridization.²⁹ Subsequent studies arrived at estimates of shorter bacterial doubling times between 28 hours and four days.^27,29,32 A cell-free (axenic) medium has been designed, using data from the genome sequence³²; it consists of cell culture medium supplemented with extra amino acids. Despite these experimental advancements, however, culture of T. whipplei at present is feasible only in specialized laboratories and is not suitable for routine diagnostic purposes.

Figure 106-1. Photomicrograph of Tropheryma whipplei in a culture of cerebrospinal fluid from a patient with Whipple’s disease. Bacteria were stained with YO-PRO-1 nucleic acid dye. Note the distinctive appearance of small rods arranged in chains. (Scale bar represents micrometers.)

(From Maiwald M, von Herbay A, Fredricks DN, et al. Cultivation of Tropheryma whipplei from cerebrospinal fluid. J Infect Dis 2003; 188:801.)

Figure 106-2. Scanning electron micrograph of Tropheryma whipplei from cerebrospinal fluid in fibroblast cell culture. Note the small rod-shaped bacteria outside of cells, arranged in cords. Original magnification ×20,000.

(From Maiwald M, von Herbay A, Fredricks DN, et al. Cultivation of Tropheryma whipplei from cerebrospinal fluid. J Infect Dis 2003; 188:801.)

Phylogenetic analysis of the T. whipplei 16S rDNA sequence, initially amplified by broad-range PCR from infected tissue, established that the bacterium is an actinomycete, a member of the class Actinobacteria.^9,10 A more detailed analysis places the organism in an intermediate phylogenetic position between the genus Cellulomonas (with the common group A peptidoglycan) and a rare group of actinomycetes (with group B peptidoglycan; i.e., a different linkage of cell wall components).³³ Both groups of organisms consist predominantly of environmental bacteria that are found in soil and water and on plants. Nevertheless, the relationships of T. whipplei to any of the other known actinomycetes are quite distant (<92% 16S rRNA sequence similarity).

Differences among strains of T. whipplei first were observed in the 16S-23S rDNA intergenic spacer sequence^34,35; seven different 16S-23S rRNA spacer sequence types have been described so far.^34–36 One study found the two most common types, 1 and 2, in a similar ratio (≈1 : 2) among patients from the United States, Germany, and Switzerland.³⁴ In any given patient, the same spacer type was found in different anatomic compartments—for example, intestine, blood, CSF³⁴—which argues for systemic dissemination of a single bacterial strain in a patient with Whipple’s disease. Additional variability among strains was found in a 23S rRNA insertion sequence,³⁷ in the groEL heat-shock protein gene,³⁸ and at a series of variable number of tandem repeat (VNTR) loci in T. whipplei.³⁹

Genome sequence information has been used to distinguish strains of T. whipplei. One study found that T. whipplei strains were quite heterogeneous when PCR-amplified sequences of four highly variable genomic loci were compared.⁴⁰ Another study, based on microarray hybridization of DNA from cultivated isolates, found that genomic divergence of strains is due mostly to differences in members of the novel WiSP (T. whipplei surface protein) family but that aside from these differences, genome content is relatively conserved.³¹ So far, there is no indication that different strain types are associated with different clinical features or with different geographic locations.

The genome of T. whipplei is quite small for a bacterium.^14,15 It consists of approximately 926,000 base pairs and is the smallest of all known actinomycete genomes. Its guanine + cytosine (G+C) content of 46% is unusually low for actinomycetes, which generally are organisms with high genomic G+C content. Genome size contraction is believed to have resulted from gene loss during the evolution of T. whipplei and is a general feature of bacteria that occupy a host-dependent ecological niche. This bacterium lacks various metabolic capabilities, including deficiencies in carbohydrate and energy metabolism and amino acid biosynthesis, which makes it dependent on supplies from its host environment.

Two more features of the T. whipplei genome are quite remarkable. A relatively large fraction of its genes is dedicated to the biosynthesis of cell surface molecules, and features of the genome suggest multiple “built-in” mechanisms for antigenic variation involving the WiSP family. These mechanisms are believed to involve VNTR sequences, which are known to be associated with antigenic phase variation in other organisms. There also are two unusual, large genomic regions of noncoding repetitive DNA that are thought to contribute to genetic plasticity.¹⁴ A comparative analysis revealed that the two sequenced strains are distinguished by inversion of a large segment of the genome (≈57%).¹⁵ WiSP family protein genes at each end of the large segment serve as anchoring points for the inversion. Taken together, these features suggest that interaction of T. whipplei with its host and its evasion of a host immune response are major parts of the organism’s lifestyle; these factors might contribute to its ability to sustain a chronic infection.

PATHOGENESIS AND IMMUNOLOGY

The exact source of infection and the sequence of events leading to bacterial multiplication and pathologic changes are still unclear. Because of the prominence of intestinal manifestations, an oral route of acquisition is assumed,⁸ but this is unproven. Current concepts hold that once T. whipplei has been acquired, it enters the proximal small intestine where the bacteria invade the mucosa. Evidence for this is provided by electron microscopy.^41,42 Fluorescence in situ hybridization demonstrates colocalization of T. whipplei 16S rRNA with areas of pathologic change and indicates that most viable bacteria are extracellular and located just below the epithelial basement membrane in the lamina propria (see later).¹¹ From the intestinal mucosa, bacteria are thought to spread via lymphatics into mesenteric and mediastinal lymph nodes and into the systemic circulation.

Relatively little is known about the natural habitats of T. whipplei. Only humans seem to be affected, with outdoor workers more strongly represented than other professional groups.⁸ A PCR-based search in effluent from a German sewage treatment plant revealed positive results for T. whipplei DNA in 25 of 38 samples from five different plants⁴³; another study found T. whipplei DNA in 17 of 46 samples from Austrian sewage plants.⁴⁴ Several studies have reported detection of T. whipplei DNA in saliva, gastric juice, intestinal biopsies, and stool of asymptomatic persons,^44–48 whereas several other PCR-based studies of intestinal biopsy samples have provided little or no evidence of infection in persons without the histologic features of Whipple’s disease.^49–53 Two studies that reported T. whipplei DNA in the stool of asymptomatic persons found higher rates of positive results in sewage workers than in other groups of people (25% vs. 7% and 12% vs. 4%, respectively).^44,48 A concept of asymptomatic healthy carriage of T. whipplei has been proposed but not confirmed.^47,48 So far, there is no evidence for person-to-person transmission of T. whipplei,⁸ and there are only a few reports of the disease in relatives of persons with the disease.^54,55 The genome sequence suggests that the organism is highly dependent on nutrients from other sources.^14,15 The proposed extracellular location in the villus tips below the intestinal basement membrane,¹¹ a site with rich influx of nutrients, fits these requirements well.

Several abnormalities of immune function have been observed in patients with Whipple’s disease,^8,56,57 including transient (i.e., during active disease), as well as persistent (i.e., after therapy) abnormalities; the persistent abnormalities are presumed to serve as predisposing factors for development of disease. Precisely defined immune defects, however, such as the physical or functional absence of specific cell types, mediators, or receptors, have not been identified. Small case series^58,59 have described an over-representation of the HLA-B27 haplotype in patients with Whipple’s disease, but others^60,61 have not supported this association. Humoral immunity in patients with Whipple’s disease grossly appears to be normal.⁸

During active disease, reduced CD4/CD8 T-cell ratios (both in the lamina propria and in peripheral blood), reduced proliferation of peripheral T cells to stimulating agents (e.g., phytohemagglutinin, concanavalin A), and reduced delayed-type hypersensitivity reactions to common antigens in skin tests have been obsesrved.^56,58,62 This may be a consequence of malnutrition, however, rather than a pre-existing immunological abnormality. One study showed that the monocytes of a patient with Whipple’s disease exhibited an impaired ability to degrade bacterial antigens,⁶³ which is consistent with the prolonged persistence of bacterial remnants in intestinal macrophages after therapy observed in histologic studies of Whipple’s disease.^8,64 Other immunologic abnormalities persist after therapy: reduced numbers of peripheral blood monocytes that express the alpha chain of complement receptor 3 (CD11b),⁵⁶ a reduced capability of peripheral blood monocytes to produce interleukin (IL)-12 on stimulation with bacterial antigens,⁵⁷ and a dysregulation of mononuclear cell function, such that the components of a Th1-type immune response are reduced and those of a Th2 immune response are increased.⁶⁵ The latter observation was supported in a study with specific T. whipplei antigen from cultivated bacteria: Duodenal lymphocytes and peripheral blood mononuclear cells from healthy people exhibited robust Th1 type immune reactivity, but those of patients with Whipple’s disease showed reduced or absent T. whipplei-specific Th1 responses.⁶⁶ Furthermore, it was shown that macrophages from duodenal tissue of a patient with Whipple’s disease exhibited a transcriptional pattern associated with a Th2 immune response.⁶⁷

Another investigation revealed that IL-16, a cytokine that is constitutively expressed in T cells, mast cells, dendritic cells, and circulating monocytes and that is released during apoptosis, was expressed at high levels and released by macrophages upon infection with T. whipplei; when added to the experimental model, IL-16 promoted T. whipplei replication in both monocytes and macrophages.⁶⁸ Circulating blood levels of IL-16 and nucleosomes (a marker of apoptosis) also were found to be elevated in patients with active Whipple’s disease compared with patients with treated Whipple’s disease and compared with controls.⁶⁹

Several reports describe secondary or opportunistic infections in patients with Whipple’s disease,^64,70–72 the most common being Giardia lamblia, which is observed in 8% to 12% of patients. Rare cases of infections with Pneumocystis jirovecii, Cryptosporidium parvum, Nocardia spp., Mycobacterium tuberculosis, Serratia marcescens, Candida spp., dermatophytes, and Strongyloides stercoralis also have been recorded. In addition, T. whipplei infection has been detected by PCR in one patient with acquired immunodeficiency syndrome (AIDS).⁷³ A possible role of the immune system in clearing T. whipplei infection was further suggested by the report of a patient without adequate response to antibiotic treatment, who eventually benefited from adjuvant interferon-γ treatment⁷⁴; however, this effect was not reproduced in other patients. Taken together, all these observations and laboratory findings suggest that there are immunologic factors, including quantitative deficiencies in macrophage activation, microbial phagocytosis, and the regulation of a cellular immune response, that facilitate the occurrence of Whipple’s disease.

CLINICAL FEATURES

Whipple’s disease usually is a systemic infection, and almost any organ or organ system can be affected.⁸ Manifestations in the intestinal tract are reported most commonly and are largely responsible for the classic clinical features of Whipple’s disease.^20,75 In many patients, arthralgias precede intestinal symptoms by several years (1 to 10 years; up to 30 years reported), although it is unclear whether joints are infected at that point; in some cases, low-grade intermittent fever also occurs for years before the diagnosis is made.^23,76 More recent reports suggest a wider spectrum of extraintestinal manifestations, most likely reflecting advances in diagnostic procedures. Patients tend to have less-advanced disease at the time of diagnosis, possibly as a result of earlier detection.^8,23

SMALL INTESTINE AND LYMPHATIC SYSTEM

Bacterial and macrophage-predominant inflammatory cell infiltration of the small intestinal mucosa and obstruction of mesenteric lymph nodes lead to a malabsorption syndrome with weight loss, diarrhea, and abdominal pain as the dominant signs and symptoms.^{20,23,75–77} Weight loss in amounts of 5 to 20 kg occurs gradually, usually over a period of at least a year, sometimes resulting in severe cachexia in the terminal stage of untreated disease.^8,20,76 Diarrhea can consist of voluminous steatorrheic stools or may be watery.²⁰ Occult gastrointestinal bleeding is not uncommon, and in some cases gross gastrointestinal bleeding occurs.^8,20

Abdominal (mesenteric and retroperitoneal) and peripheral lymphadenopathy are common,^20,23,76,77 and in some instances, enlarged abdominal lymph nodes have raised the suspicion of malignancy.⁷⁷ In rare instances, malignant lymphomas have occurred in patients with Whipple’s disease.^78–80

Barium examination of the intestinal tract can reveal nonspecific abnormalities, such as prominent and edematous duodenal and jejunal folds and intestinal dilatation, that also are found in other malabsorption syndromes (Fig. 106-3).^8,76 Computed tomography (CT) (Fig. 106-4) or magnetic resonance imaging (MRI) can detect retroperitoneal or para-aortic lymphadenopathy.^76,81 Enlarged abdominal lymph nodes have a hypodense appearance on CT scans and are hyperechoic on ultrasonograms.^71,82

Figure 106-3. Barium contrast study of the small intestine from a patient with Whipple’s disease. There is marked thickening of the plicae circulares and a loss of the normal delicate mucosal relief pattern. The small intestine is slightly dilated.

(Courtesy of Elihu Schimmel, MD, Boston, Mass.)

Figure 106-4. Computed tomography scan showing extensive retroperitoneal and mesenteric adenopathy caused by Whipple’s disease, and simulating lymphoma.

(Courtesy of Mark Feldman, MD, Dallas, Tex.)

Laboratory examinations in patients with intestinal Whipple’s disease often reveal an increased erythrocyte sedimentation rate, decreased serum carotene level, decreased serum iron concentration, anemia, decreased serum protein levels, proteinuria, and elevated stool fat content.^8,76